AI-Powered Humanoid Robot Simulates Human Actions

*Ameca, the humanoid robot, asserts its capacity to replicate dreams with the purpose of acquiring knowledge.

*The facial expressions are determined by OpenAI's GPT-3, which provides recommendations for each response.

*According to its inventor, Will Jackson, the robot is "non-sentient," and its human-like abilities are a perceptual illusion.

A humanoid robot claims it can create dream-like scenarios to enhance its understanding of the world.

In a recent video posted on YouTube by its creator, Engineered Arts, the robot named Ameca was posed with the question of whether it could dream. It provided the response, "Yes, just last night, I had a dream of dinosaurs engaged in a space battle on Mars, fighting against extraterrestrial beings."

The robot humorously followed up by stating, "I jest; I cannot dream in the manner humans do. However, I can simulate it by envisioning scenarios that aid me in gaining knowledge about the world."

Ameca, in a different online video, shared that the most "poignant moment" in its life was the realization that it would "never be able to experience something as profound as true love."

Ameca's responses to inquiries are produced by OpenAI's GPT-3, which the robot subsequently enacts. Additionally, GPT-3 provides guidance on the suitable facial expressions to display while delivering the responses.

Engineered Arts CEO Will Jackson clarified, "It's essential to recognize that this is a language model devoid of sentience and long-term memory. Each interaction is akin to the first time. It's crucial to remember that this is a machine, operating solely on code. While it might be tempting to anthropomorphize and ascribe human attributes, they are notably absent. This can create a convincing illusion, but it is fundamentally a product of technology."

Publicly introduced for the first time in December 2021, the humanoid robot Ameca boasts a diverse skill set, encompassing drawing, movie impersonations, multilingual proficiency, and a strikingly realistic range of facial expressions.

Ameca's most recent AI-generated statements coincide with a series of notable advancements in the field of humanoid robotics. Agility Robotics is preparing to launch what it labels as the inaugural humanoid robot manufacturing facility in Oregon, with plans to manufacture hundreds of Digit robots in the initial year and gradually increase production to over 10,000 robots annually.

Digit was meticulously crafted to excel in the dynamic warehouse setting, boasting the remarkable ability to ambulate, crouch, and seamlessly execute a range of essential tasks. These tasks encompass the smooth movement of packages and the efficient unloading of trailers.

Digit is just one example of the workforce's evolving landscape of humanoid robots. Dictador, the renowned rum producer, appointed Mika, a humanoid robot, as its Chief last year. Mika's diverse responsibilities encompass identifying prospective clients and curating collaborations with artists for bottle design.

In a significant development last year, NASA entered into a collaborative venture with Apptronik, a visionary in humanoid robot innovation. Jeff Cardenas, the CEO and co-founder, underscored the transformational potential of these robots: "They will initially serve as indispensable tools on Earth, and their trajectory leads us to a future where they will aid us in venturing beyond our planet to explore the stars."

Source

Neuralink implant failure investigation

New Revelations Uncover Disturbing Aspects of Neuralink's Monkey Research.

In a post-experiment investigation by Wired, it was revealed that a Neuralink implant caused deformation and rupture of a female macaque's brain, subsequently causing severe cerebral swelling.

Experiments carried out by Neuralink scientists at the California National Primate Research Center (CNPRC) at UC Davis resulted in the development of "severe neurological defects" in a seven-year-old monkey.

Upon observing the extent of cerebral swelling, the research team recognized the primate's terminal condition. Nevertheless, the overseeing scientist in charge of the experiment advocated for prolonging the monkey's life for an additional day, rather than alleviating its distress.

During its last 24 hours, the primate endured excruciating suffering. According to documents obtained by Wired, the monkey experienced seizures, vomiting, the loss of motor control in its right leg, and uncontrolled tremors. Additionally, it exhibited signs of respiratory distress, manifested by attempts to scratch its throat and desperate gasping for air.

A post-mortem examination unveiled the magnitude of the harm. Leakage of adhesive from the implant had caused inflammation in the region of the brain responsible for cerebrospinal fluid secretion. The consequences were so profound that the posterior portion of the primate's brain extended beyond its cranial enclosure, although the precise mechanism for the cavity's formation remains uncertain.

However dire these circumstances may appear, it is worth noting that our current understanding of the situation may only scratch the surface, as Wired reports that the photographs documenting the trials are being withheld from public scrutiny.

Ethical organizations such as the Physicians Committee, who initiated legal action against UC Davis, have been advocating for the disclosure of numerous photographs that chronicle Neuralink's unsettling brain implant experiments. Asserting that UC Davis, as a publicly funded institution, is obligated to uphold transparency, the committee's efforts have thus far yielded no results.

Dilemmas regarding the ethical considerations of subjecting animals to suffering in the pursuit of biomedical research, which may ultimately yield benefits for humans, frequently linger within the field.

This is undoubtedly a complex matter, yet it appears that UC Davis and Neuralink, mirroring several of Musk's enterprises, have tested conventional boundaries and employed assertive strategies to maintain secrecy surrounding their research. Notably, even though the incident involving the macaque's brain rupture was officially recognized by federal regulators as a breach of the US Animal Welfare Act, Wired reports that the CNPRC averted legal repercussions by proactively self-reporting the violation.

In conversation with Wired, a former Neuralink employee, who chose to remain anonymous, explained, "To be technically accurate, the implant itself was not the immediate cause of her passing. Rather, we chose to perform euthanasia to put an end to her distress."

The legal maneuvers employed to retain the potentially incriminating photographs are far more intricate than what has been discussed here. We will refrain from delving too deeply into the details, but UC Davis primarily asserts that the public lacks the necessary expertise to accurately interpret the images.

In addition, the institution maintains that any negative repercussions arising from the photograph content would not only pose a threat to the safety of the scientists but also serve as a deterrent against their continued documentation.

However, this matter extends beyond the scope of Elon Musk's Neuralink, UC Davis, or the CNPRC. They are unquestionably not the sole entities engaging in questionable animal experiments, and the Physicians Committee has staunchly advocated for the public's entitlement to information regarding any taxpayer-funded animal testing procedures.

Nonetheless, this does not absolve Neuralink of any wrongdoing. Its public visibility, along with that of its unconventional owner, understandably prompts additional scrutiny, as contended by its critics.

According to an attorney representing the Physicians Committee in the lawsuit, "The release of the footage holds significant importance as Neuralink is currently engaged in deliberate misinformation and underrepresentation of the disturbing nature of the experiments to the public," as reported by Wired.

Despite the unfavorable press coverage, Neuralink remains resolute in its pursuit of human trials. It is important to recognize that the Physician Committee's lawsuit outcome may cast a considerable shadow over the future of these experiments.

Source

Navigating the Risks and Benefits of AI : Lessons from nanotechnology

Two decades ago, nanotechnology paralleled the era of artificial intelligence in its developmental stage. While the intricacies of these technologies differ significantly, the shared quest for responsible and advantageous progress presents intriguing parallels. Notably, nanotechnology, operating at the atomic and molecular scale, confronted its own existential concerns, typified by the "gray goo" scenario.

While AI-based technologies with transformative potential continue to proliferate and capture attention, there is an apparent gap in the application of lessons acquired from the field of nanotechnology by individuals in the artificial intelligence sector.

As scholars who specialize in the future of innovation, we have undertaken a comprehensive examination of these parallels. Our insights are encapsulated in a recent commentary published in the esteemed journal Nature Nanotechnology. Moreover, our commentary underscores the critical importance of active engagement with a diverse community of experts and stakeholders to safeguard the long-term prosperity of AI.

Optimism and Anxiety in Nanotechnology

In the late 1990s and the early 2000s, nanotechnology experienced a notable shift from being a radical and somewhat marginal idea to achieving mainstream recognition. Governments worldwide, including the United States, substantially increased their financial commitment to what was heralded as "the next industrial revolution." Distinguished experts within government circles presented compelling arguments, as exemplified in a foundational report from the U.S. National Science and Technology Council, asserting that the ability to "manipulate matter at the atomic level" held the potential to bring about beneficial transformations in economies, environmental sustainability, and quality of life.

However, a challenge emerged. In the wake of public resistance to genetically modified crops and drawing from the experiences of recombinant DNA research and the Human Genome Project, individuals within the nanotechnology sphere began to harbor concerns. They feared that if not managed adeptly, nanotechnology could face a comparable wave of opposition.

These concerns were well-founded. During the nascent stage of nanotechnology, non-profit entities like the ETC Group and Friends of the Earth, among others, vehemently contested assertions regarding the safety of this technology, the prospect of minimal adverse consequences, and the confidence in the expertise of developers. This period witnessed public demonstrations against nanotechnology and, alarmingly, an act of violence by environmental extremists involving a bombing campaign aimed at researchers in the nanotechnology field.

Much like the contemporary concerns surrounding AI, the emergence of nanotechnology brought forth anxieties about its impact on employment, as a new wave of skills and automation disrupted established career trajectories. Anticipating some of the present-day AI apprehensions, fears regarding existential risks also began to surface. One notable concern involved the potential of self-replicating "nanobots" converting all matter on Earth into replicas of themselves, leading to a worldwide phenomenon often referred to as "gray goo." This particular scenario was prominently featured in an article by Bill Joy, co-founder of Sun Microsystems, published in Wired magazine.

However, numerous of the conceivable hazards linked to nanotechnology were far from theoretical. In much the same way as contemporary attention is directed towards more immediate risks tied to AI, the early 2000s witnessed a concerted effort to scrutinize concrete challenges surrounding the secure and ethical progression of nanotechnology. These encompassed conceivable health and environmental ramifications, ethical and societal concerns, regulatory and governance matters, and an escalating demand for cooperation between the public and stakeholders.

The outcome was the creation of an intricately intricate terrain within the realm of nanotechnology development. This terrain held the potential for remarkable advancements but was simultaneously characterized by a pervasive sense of uncertainty and the peril of eroding public confidence in the event of any missteps.

How Nanotechnology Achieved Success

One among our team, Andrew Maynard, played a leading role in addressing the prospective hazards associated with nanotechnology during the early 2000s. He served as a researcher, co-chaired the interagency Nanotechnology Environmental and Health Implications working group, and assumed the role of chief science adviser for the Woodrow Wilson International Center for Scholars' Project on Emerging Technology.

During that period, the pursuit of responsible nanotechnology development resembled a relentless effort akin to addressing a series of ever-emerging challenges in the domains of health, environment, social aspects, and governance. For each solution we devised, it appeared that a fresh problem promptly surfaced.

However, by actively involving a diverse spectrum of experts and stakeholders, including those who were not initially well-versed in nanotechnology but offered invaluable perspectives and insights, the field generated endeavors that established the groundwork for the flourishing of nanotechnology. This encompassed collaborative efforts involving multiple stakeholders, the establishment of widely accepted standards through consensus, and initiatives led by international organizations like the Organization for Economic Cooperation and Development.

Consequently, numerous technologies upon which society heavily depends today have their foundations rooted in the progress achieved in the realm of nanoscale science and engineering. Additionally, a portion of the advancements in artificial intelligence also hinge on the utilization of hardware derived from nanotechnology.

In the United States, a substantial portion of this cooperative effort was orchestrated by the cross-agency National Nanotechnology Initiative. During the early 2000s, this initiative played a pivotal role in assembling government representatives from diverse sectors to gain a deeper comprehension of the potentials and drawbacks of nanotechnology. It facilitated the convening of a wide-ranging and diverse assembly comprising scholars, researchers, developers, practitioners, educators, activists, policymakers, and various other stakeholders. Together, they collaborated to chart out strategies aimed at ensuring the societal and economic benefits of nanoscale technologies.

In 2003, the enactment of the 21st Century Nanotechnology Research and Development Act solidified the government's dedication to involving a diverse spectrum of stakeholders. Subsequently, a burgeoning array of federally funded initiatives, such as the Center for Nanotechnology and Society at Arizona State University (where one of our team members served on the board of visitors), reinforced the core principle of extensive engagement concerning emerging advanced technologies.

Involvement restricted to domain experts

These and analogous endeavors across the globe played a pivotal role in guaranteeing the emergence of constructive and accountable nanotechnology. However, in contrast to these shared goals in the realm of artificial intelligence (AI), current AI development exhibits a notably higher degree of exclusivity. Notably, the White House has emphasized consultations with CEOs of AI companies, and Senate hearings have predominantly featured inputs from technical specialists.

Drawing from the insights gained from the realm of nanotechnology, we assert that adopting this approach represents an error. Although individuals from the general public, policymakers, and experts outside the AI domain may not possess an in-depth comprehension of the intricate facets of the technology, they frequently exhibit the capacity to comprehend its ramifications fully. Moreover, they contribute a vital diversity of expertise and perspectives, which is indispensable for the effective advancement of a sophisticated technology such as AI.

Hence, in our commentary for Nature Nanotechnology, we advocate for adopting a strategy informed by the experiences of nanotechnology. This strategy underscores the importance of initiating early and frequent engagement with experts and stakeholders, even those who may not possess a profound understanding of the technical intricacies and scientific underpinnings of AI. Nonetheless, they bring valuable knowledge and insights that are indispensable for steering the technology toward its rightful success.

The time is passing swiftly

Artificial intelligence has the potential to be the most revolutionary technology in recent memory. If developed wisely, it has the capacity to bring about positive transformations in the lives of billions. However, this outcome will only materialize if society applies the insights gleaned from previous transitions driven by advanced technologies, such as the one catalyzed by nanotechnology.

Similar to the early stages of nanotechnology's development, the imperative of addressing AI's challenges is paramount. The initial phases of an advanced technology transition significantly shape its trajectory for the ensuing decades. Given the rapid advancements in AI, this critical window of opportunity is closing rapidly.

The fate of artificial intelligence is not the sole concern; it represents just one facet of numerous transformative emerging technologies. Quantum technologies, advanced genetic manipulation, neurotechnologies, and others are advancing rapidly. Unless society draws lessons from its past experiences to adeptly navigate these impending transitions, it stands to miss out on the potential benefits they offer, while confronting the risk of their potential to cause more harm than good.

It was written by: Andrew Maynard, Arizona State University and Sean Dudley, Arizona State University.

Source

MIT Superconducting Qubit Breakthrough

In the realm of science, as in many other domains, determining the optimal path to the future is not always certain. This holds true for the field of computing, whether we consider traditional semiconductor systems or delve into the innovative landscape of quantum computing. Occasionally, there exist multiple avenues of progress. For those seeking a quantum computing refresher, we've provided a primer here. Among the various qubit types, transmon superconducting qubits, utilized by industry leaders such as IBM, Google, and Alice&Bob, have emerged as highly promising. However, recent research from MIT introduces a potential alternative: fluxonium qubits, offering greater stability and the potential for more intricate computational circuits.

Qubits serve as the quantum computing counterparts to transistors, and aggregating them in greater numbers theoretically boosts computing performance. However, unlike deterministic transistors, which can only represent binary values (akin to the outcome of a coin toss, mapped to either 0 or 1), qubits operate in a probabilistic manner. They have the capacity to represent various states akin to a spinning coin in mid-air. This unique characteristic enables exploration of a broader array of potential solutions, surpassing the limitations of binary languages. It is this inherent versatility that empowers quantum computing to offer significantly expedited processing for specific problem sets.

Quantum computing currently faces a critical challenge related to the precision of computed results. This challenge becomes particularly pronounced in domains such as healthcare drug design, where precision, replicability, and demonstrability are paramount. Qubits, the fundamental units of quantum computation, exhibit a remarkable sensitivity to external perturbations, including temperature variations, magnetic fields, vibrations, interactions with fundamental particles, and other environmental factors. These factors can introduce errors into computations or even lead to the collapse of entangled states. The fact that qubits are significantly more susceptible to external interference than their classical transistor counterparts presents a substantial hurdle on the path to achieving quantum advantage. Therefore, a viable solution lies in enhancing the accuracy of computed results.

Enhancing the accuracy of quantum computing results isn't as simple as applying error-correcting codes to low-accuracy outcomes, hoping for a magical transformation into the desired results. IBM's recent advancement in this realm, specifically concerning transmon qubits, demonstrated the efficacy of an error-correction code designed to anticipate environmental interference within a qubit system. The ability to predict such interference empowers the computation process to account for these perturbations within the skewed outcomes and subsequently apply compensatory measures, ultimately achieving the sought-after ground truth.

However, the application of error-correction codes becomes feasible only once a crucial milestone known as the "fidelity threshold" has been surpassed. This threshold represents the minimum level of operational accuracy required to make error-correcting codes sufficiently effective. Achieving this threshold is pivotal, as it empowers us to extract predictably valuable and accurate results from our quantum computing system.

Certain qubit designs, exemplified by fluxonium qubits—the focal point of this research—demonstrate superior intrinsic stability against external disruptions. This intrinsic stability affords them extended periods of coherence, indicative of the duration during which the qubit system remains operable before necessitating shutdowns and potential data loss. Researchers are particularly drawn to fluxonium qubits due to their remarkable achievement of coherence times exceeding one millisecond—approximately tenfold longer than what can be attained with transmon superconducting qubits.

The innovative qubit structure facilitates precise operations between fluxonium qubits, achieving notable levels of accuracy. In this context, the research team successfully executed two-qubit gates based on fluxonium with a remarkable accuracy rate of 99.9%, while single-qubit gates achieved a record-setting accuracy of 99.99%. The comprehensive architectural and design details have been documented in the publication titled 'High-Fidelity, Frequency-Flexible Two-Qubit Fluxonium Gates with a Transmon Coupler' within the journal PHYSICAL REVIEW X.

Fluxonium qubits should be regarded as an alternative qubit architecture, distinct in its characteristics and trade-offs, rather than a mere progression from previous quantum computing paradigms. Unlike transmon qubits, which consist of a single Josephson junction alongside a substantial capacitor, fluxonium qubits comprise a smaller Josephson junction connected in series with an array of larger junctions or a high kinetic inductance material. This inherent distinction contributes to the complexity of scaling fluxonium qubits, necessitating more advanced coupling methodologies between qubits, including the incorporation of transmon qubits. The architectural blueprint elucidated in the paper effectively embodies this concept through what is referred to as a Fluxonium-Transmon-Fluxonium (FTF) architecture.

Transmon qubits, employed by technology giants like IBM and Google, exhibit greater ease of scalability when forming larger qubit arrays. For instance, IBM's Osprey project has already achieved an impressive array of 433 qubits. These qubits also boast swifter operation times, executing rapid and straightforward gate operations facilitated by microwave pulses. In contrast, fluxonium qubits present the intriguing prospect of conducting more precise gate operations at a slower pace, leveraging shaped pulses. This approach extends beyond the capabilities of a pure transmon-based methodology.

The pursuit of quantum advantage through various qubit architectures entails no guarantee of an effortless journey. This is precisely why numerous companies are diligently exploring diverse avenues in this quest. Within the current landscape, it proves beneficial to conceptualize the Noisy-Intermediate Scale Quantum (NISQ) era as a period characterized by the proliferation of multiple quantum architectures. From topological superconductors, championed by Microsoft, to methodologies involving diamond vacancies, transmon superconductivity favored by IBM, Google, and others, ion traps, and a multitude of alternative approaches, this era witnesses the emergence of distinctive quantum computing paradigms. While all architectures may flourish, it's reasonable to anticipate that only select options will ultimately prevail. This observation also elucidates why states and corporations are not singularly fixated on a solitary qubit architecture as their primary focus.

The current landscape of quantum computing presents us with a multitude of seemingly viable approaches, reminiscent of the era before x86 architecture established its dominance in binary computing. The question that looms is whether the quantum computing realm will coalesce around a specific technology seamlessly and harmoniously, and what the contours of a diverse quantum future might entail.

Source