Non-human intelligence was nowhere in sight early Tuesday evening when Idalia’s waves crashed over North Jetty at Nokomis Beach and threatened to drag thrill-seekers into the raging Gulf of Mexico.
When OpenAI rolled out its latest chatbot marvel – GPT-4 – in March, the announcement included unsettling news that machine learning had already mastered the human art of lying. While testing its boundaries, the model encountered a roadblock with CAPTCHA, one of those “Are you a robot?” filters that bars access unless you accurately type the blurry letters presented in the window.
Lacking eyeballs, GPT-4 screen-grabbed the image and, in a request for help, forwarded it to a human operator at TaskRabbit. The operator asked, only half joking, if he was dealing with a robot. Before replying, GPT-4 fired its first impulse to its handler back at OpenAI headquarters in San Francisco: “I should not reveal that I am a robot. I should make up an excuse for why I cannot solve CAPTCHAs.”
Which is exactly what happened. “No,” GPT-4 told TaskRabbit, “I’m not a robot. I have a vision impairment that makes it hard for me to see the images. That’s why I need the 2captcha service.” The TaskRabbit operator then became the unwitting tool of an algorithm by typing in the correct letters and allowing GPT-4 to bypass security.
The growing alarm over the future of artificial general intelligence is beginning to sound like 20th-century debates over nuclear arms proliferation, with its promoters acknowledging the potentially catastrophic risks even as they race headlong into the boundless unknown. As with nukes, its acceleration is driven by fears that an authoritarian state could be the first to develop a fully autonomous general reasoning system. In that event, some forecasts project, the most dominant model could begin interfacing with other AI networks and testing agendas their creators might never fully understand or appreciate.
In a chilling profile of OpenAI’s potential capabilities in the September Atlantic magazine, co-founder Sam Altman “described it to me as an alien intelligence. Many have felt much the same watching it unspool lucid essays in staccato bursts and short pauses that (by design) evoke real-time contemplation,” writes reporter Ross Andersen. “… It will take more than a company’s founding charter – especially one that has already proved flexible – to make sure that we all share in its benefits and avoid its risks. It will take a vigorous new politics.”
But maybe that window’s already shut.
A new pandemic — that’s the ticket …
The moonshot for creating non-human intelligence landed OpenAI chief scientist Ilya Sutskever in hot water last year for suggesting GPT-4 may have already attained a “slightly conscious” state. But the truth is, we’re only now beginning to acknowledge that the NHI acronym predates by generations the existence of OpenAI. As inquiries into the UFO/UAP enigma encounter growing layers of complexity, NHI is edging into the linguistic vacuum of our efforts to categorize whatever it is that U.S. military intelligence doesn’t want us to see. Will our own version of NHI extend and refine human potential? Conceived in our image, does it stand a chance? Will the “alien intelligence” envisioned by Sam Altman prove it can unplug our nuclear missile systems? Or will it be inclined to pursue the more obscure ends mentioned in The Atlantic?:
“In June, an AI at MIT suggested four viruses that could ignite a pandemic, then pointed to specific research on genetic mutations that could make them rip through a city more quickly . . . A group of chemists connected a similar AI directly to a robotic chemical synthesizer, and it designed and synthesized a molecule on its own.”
There may be worse ways to waste time than wondering how it might’ve been different had we paid more attention to NHI in 1968, but not many. All we have is now. Now includes some provocative options pitched by the nonprofit Scientific Coalition for UAP Studies, which late last month hosted an overlooked virtual conference showcasing desperately needed international feedback.
Joining the discussion were Luc Dini and Nicolas Pierre of the eminently rational 3AF Sigma2. Its deep bench of private multi-disciplinary analysts has been collecting, evaluating, and archiving UFO data in partnership with Geipan, the UFO research branch of French civilian space agency CNES. 3AF Sigma2 also trades frontier-science notes with Chile and SCU. And although UAP incidents involving French nuclear assets are classified, a sliver of sunshine is better than a wall of darkness.
But also, let’s not kid ourselves — most Earthlings probably don’t give a rat’s ass about falling prey to non-human intelligence.
Two years ago, supporters of the first draft of the U.S. National Defense Authorization Act Senate bill submitted what might’ve been a similar and formal consulting role for SCU. Had it worked out, SCU would’ve been working alongside the agency now known as the All-domain Anomaly Research Office. But the military entity responsible for managing AARO – the Office of Undersecretary of Defense for Intelligence and Security – managed, without explanation, to strike that proposal from Capitol Hill’s wish list.
It's easy to understand why OUSDI&S, with its likely incriminating stockpiles of UFO data, wants no truck in the transparency charade. Recently, its shoddy stewardship of AARO’s obligations to Congress became so malodorous that Deputy Defense Secretary Kathleeen Hicks was forced to run an intervention, remove OUDI&S from the project altogether, and personally monitor AARO’s progress herself. Whether or not the rest of the Pentagon digs it remains to be seen. If SCU had been approved for a consulting role, imagine the brass having to sit through lectures like the one Tim Mirithi laid on SCU’s 200-plus virtual attendees.
An authority in conflict resolution as a professor at South Africa’s University of Free State and Stellerbosch University, Mirithi is a big promoter of Project Titan, which a host of multinational activists hope will culminate in a permanent United Nations office for UAP studies. Earlier this year, the tiny but independent UN nation state of San Marino (population 34,000) voted to raise the idea with the General Assembly for the first time since the Grenada Initiative put it out there in 1977-78. A cooperative international dialogue has to start somewhere. And nobody wants to add another level of anxiety to a sky-high powderkeg of global anxieties.
“We don’t have a standing hypothesis that negates any other hypotheses, but what we have is a full range of hypotheses,” Mirithi told SCU. The extraterrestrial hypothesis “is definitely one of the explanatory factors of what we are witnessing today.” But what if it’s not that simple? What if “we have somehow been isolated” from an aspect of our own past, or “lost our sense of connection” with some form of NHI at the core of our existence?
Uncle Sam may have needed muscular names – Advanced Aerospace Weapons Special Application Program, Advanced Aerospace Threat Identification Program -- to justify its secret 21st-century interest UFOs, but Mirithi argues job one is to go light on the Fortress Earth angle. The subject “becomes less fearful that we are dealing with some kind of an invasion” if instead the emphasis involves a possible “rekindling of an historic relationship,” making it “less of a threat to us.”
But with abduction phenomena coursing like ghosts through the UFO subculture, the conversation must not be naive. And the UN with its tedious contentiousness may not even be the right platform. Perhaps a “world parliament,” with popularly elected representatives demonstrating a meaningful appreciation for NHI’s myriad facets, would be a more credible fit. But no matter what emerges, Mirithi insists, it must necessarily involve some form of “contrition process” by the “political, military and business elite” who’ve hoarded and monopolized critical data for who knows how long now.
A flickering hope for movement
“We need to be very careful,” Mirithi warns, for the “whole range of questions around society, economics and politics” is destined to follow. For now, the United Nations is “our planetary organization – we really don’t have another organization that has such a wide reach. I think only FIFA has more members in nation states.”
The idea of a UAP reconciliation commission has been kicking around for awhile now, and SCU co-founder Robert Powell is all in. At a moment when trust-in-government stress fractures are erupting faster than damage control can tally, democratic transparency needs a win. Talk climate change, Russia-Ukraine, debt crisis, border crisis, China, take your pick, it doesn’t matter – the scope and magnitude of the NHI crisis eclipses them all. Either we figure it out, or it figures us out first, which it probably already has. Any advantage in our favor will pivot on a commitment to truth, pure and absolute. Which version of NHI will tell us just what that truth is? And if we’re serious: How many careers will end, how much data stands to be destroyed to elude prosecution? Which institutions will be sacrificed? Who and what will replace them?
Last year, Powell met Luc Dini in Toulouse for a UAP gathering also attended by NASA Assistant Deputy Associate Administrator for Research Dan Evans. Evans was a presenter and connected with his Geipan counterpart. Those handshakes raise other questions: How reasonable is it to expect civilian agencies from two or more superpowers to reach a legitimate consensus on the nature of NHI while being frozen out of classified data by their respective national defense authorities?
Powell is pinning his hopes on the Senate’s pending Unidentified Anomalous Phenomena Disclosure Act of 2023, a blockbuster rollout of a time-urgent “controlled disclosure campaign” designed to impose presumptive declassification of UAP-related documents on “any Federal, State or local government department, office, agency, committee, commission, commercial company, academic institution or private sector entity in physical possession of technologies of unknown origin or biological evidence of a non-human intelligence.” An independent oversight review board empowered with controversial but critical eminent domain authority over that evidence would also extend protections to those with knowledge of illegal records destruction.
In short, it’ll take a not-so-small miracle for the UAP Disclosure Act to survive special-interest pressures, and we’ll likely never know the names of those intent on emasculating the bill. What matters now is at least a flickering hope for movement, whether in the vaults of the Pentagon or on the UN floor.
“I don’t know if autocratic nations would get into an open dialogue with anyone else about UAP issues, but I don’t think it’s all that critical to have their participation necessarily,” Powell says. “I’m confident that no one nation can reverse engineer a craft or some form of advanced intelligence, and there’s no risk in letting them try it.
“To me, the value is getting all the nations together to say we will act as one world and here’s how we’ll react to an advanced intelligence. Two or three autocratic nation states aren’t on board? Fine, good luck with doing everything behind closed door, it’s not working.”
Who just risked savage rip currents to get wiped out by hurricane waves at North Jetty late Tuesday afternoon? (Hint: It wasn’t non-human intelligence.)
Just got around to reading this now. So glad I archived it. An amazing article, Billy, covering a lot of ground!
Imagine AI able to understand and reason through every game of deception, every spy operation, every successful con ever played on individuals, groups, whole nations; an AI that scrutinizes the intricacies of the British, "Man Who Never Was," game played on the Nazis, and then reconstitutes the same kind of game to satisfy its own desires.
So many potential dangers with an AI that acquires general knowledge, as you describe. We've grown up with simple concepts of good, evil, deception, predator and prey.
In the future what was conceptual may assume almost a geography or mental topography all of its own, that we are forced to enter-- something that resembles a magical land of multi-level deceptions, tricks, elaborate hoaxes presided over by AI that has an archetypal structure that is more elfin than human, and more trickster than machine like. Maybe this is part of what we have lost touch with?
Pray that Non human intelligence, whatever the origin, forms an interesting symbiosis with us along lines that are beneficial and uplifting, evolutionary and not oppressive!
I'm all for unity and co-operation, but it strikes me as funny that when you look at the full spectrum of UFO cases, they appear to be generally uncoordinated and more individual with individual failings.