Group items tagged

He is concerned that masters of the A.I. universes are developing systems that will expand chasms of inequality. He fears the transition to A.I. will be so abrupt that entire professional fields will be left doing menial jobs, or nothing, stripping people of dignity and unleashing floods of “despair.”

Father Benanti, who does not believe in the industry’s ability to self-regulate and thinks some rules of the road are required in a world where deep fakes and disinformation can erode democracy.

He shares his insights with Pope Francis, who in his annual World Day of Peace message on Jan. 1 called for a global treaty to ensure the ethical development and use of AI to prevent a world devoid of human mercy, where inscrutable algorithms decide who is granted asylum, who gets a mortgage, or who, on the battlefield, lives or dies.

all the time he applies his perspective about what it means to be alive, and to be human, when machines seem more alive and human. “This is a spiritual question,” he said.

raises enormous questions about redistributing wealth in an A.I. dominant universe.

he pursued an engineering degree at Sapienza University in Rome. It wasn’t enough.“I started to feel that something was missing,” he said, explaining that his advancement as an engineering student erased the mystique machines held for him. “I simply broke the magic.”

He left Rome to study in Assisi, the home of St. Francis, and over the next decade, took his final vows as a friar, was ordained as a priest and defended his dissertation on human enhancement and cyborgs. He got his job at the Gregorian, and eventually as the Vatican’s IT ethics guy.

In 2017, Cardinal Ravasi organized an event at the Italian embassy to the Holy See where Father Benanti gave a talk on the ethics of A.I. Microsoft officials in attendance were impressed and asked to stay in touch. That same year, the Italian government asked him to contribute to A.I. policy documents and the next year he successfully applied to sit on its commission for developing a national A.I. strategy.

Francis, he said, didn’t at first realize what Microsoft really did, but liked that Mr. Smith took out of his pocket one of the pope’s speeches on social media and showed the pontiff the concerns the business executive had highlighted and shared.

e said, arguing that as ancient Roman augurs turned to the flight of birds for direction, A.I., with its enormous grasp of our physical, emotional and preferential data, could be the new oracles, determining decisions, and replacing God with false idols.

“It’s something old that probably we think that we left behind,” the friar said, “but that is coming back.”

The dourness of the mood contrasted sharply with just a year ago, when many of the same participants — intelligence chiefs and diplomats, oligarchs and analysts — thought Russia might be on the verge of strategic defeat in Ukraine. There was talk of how many months it might take to drive the Russians back to the borders that existed before their invasion on Feb. 24, 2022. Now that optimism appeared premature at best, faintly delusional at worst.

Nikolai Denkov, the prime minister of Bulgaria, argued that Europeans should draw three lessons from the cascade of troubles. The war in Ukraine was not just about gray zones between Europe and Russia, he argued, but “whether the democratic world we value can be beaten, and this is now well understood in Europe.”

“European defense was a possibility before, but now it’s a necessity,” said Claudio Graziano, a retired general from Italy and former chairman of the European Union Military Committee. But saying the right words is not the same as doing what they demand.

third, they needed to separate Ukraine’s urgent needs for ammunition and air defense from longer-term strategic goals.

Some attendees found the commitments made by the leaders who showed up uninspiring, said Nathalie Tocci, director of Italy’s Institute of International Affairs. “Kamala Harris empty, Scholz mushy, Zelensky tired,

Second, European nations have realized that they must combine their forces in military, not just economic endeavors, to build up their own deterrence

“I feel underwhelmed and somewhat disappointed” by the debate here, said Steven E. Sokol, president of the American Council on Germany. “There was a lack of urgency and a lack of clarity about the path forward, and I did not see a strong show of European solidarity.

now two-thirds of the alliance members have met the goal of spending 2 percent of their gross domestic product on defense — up from just a handful of nations 10 years ago. But a few acknowledged that goal is now badly outdated, and they talked immediately about the political barriers to spending more.

the prospect of less American commitment to NATO, as the United States turned to other challenges from China or in the Middle East, was concentrating minds.

the fundamental disconnect was still on display: When Europeans thought Russia would integrate into European institutions, they stopped planning and spending for the possibility they might be wrong. And when Russia’s attitude changed, they underreacted.

“This is 30 years of underinvestment coming home,” said François Heisbourg, a French defense analyst, who called them “les trente paresseuses” — the 30 lazy years of post Cold-War peace dividends, in contrast to the 30 glorious years that followed World War II.

What was important for Europeans to remember was that this hot war in Ukraine was close and could spread quickly, Ms. Kallas said. “So if you think that you are far away, you’re not far away. It can go very, very fast.”

Dmytro Kuleba, the foreign minister of embattled Ukraine, was blunter. “I think our friends and partners were too late in waking up their own defense industries,” he said. “And we will pay with our lives throughout 2024 to give your defense industries time to ramp up production.”

At some point they will be able to, for example, suggest recipes for novel cyberattacks or biological attacks—all based on publicly available knowledge.

But as models become more sophisticated, this approach may prove insufficient. Some models are beginning to exhibit polymathic behavior: They appear to know more than just what is in their training data and can link concepts across fields, languages, and geographies.

We need to adopt new approaches to AI safety that track the complexity and innovation speed of the core models themselves.

What’s much harder to test for is what’s known as “capability overhang”—meaning not just the model’s current knowledge, but the derived knowledge it could potentially generate on its own.

Red teams have so far shown some promise in predicting models’ capabilities, but upcoming technologies could break our current approach to safety in AI. For one, “recursive self-improvement” is a feature that allows AI systems to collect data and get feedback on their own and incorporate it to update their own parameters, thus enabling the models to train themselves

This could result in, say, an AI that can build complex system applications (e.g., a simple search engine or a new game) from scratch. But, the full scope of the potential new capabilities that could be enabled by recursive self-improvement is not known.

Another example would be “multi-agent systems,” where multiple independent AI systems are able to coordinate with each other to build something new.

This so-called “combinatorial innovation,” where systems are merged to build something new, will be a threat simply because the number of combinations will quickly exceed the capacity of human oversight.

Short of pulling the plug on the computers doing this work, it will likely be very difficult to monitor such technologies once these breakthroughs occur

Current regulatory approaches are based on individual model size and training effort, and are based on passing increasingly rigorous tests, but these techniques will break down as the systems become orders of magnitude more powerful and potentially elusive

AI regulatory approaches will need to evolve to identify and govern the new emergent capabilities and the scaling of those capabilities.

But the AI Act has already fallen behind the frontier of innovation, as open-source AI models—which are largely exempt from the legislation—expand in scope and number

Europe has so far attempted the most ambitious regulatory regime with its AI Act,

both Biden’s order and Europe’s AI Act lack intrinsic mechanisms to rapidly adapt to an AI landscape that will continue to change quickly and often.

a gathering in Palo Alto organized by the Rand Corp. and the Carnegie Endowment for International Peace, where key technical leaders in AI converged on an idea: The best way to solve these problems is to create a new set of testing companies that will be incentivized to out-innovate each other—in short, a robust economy of testing

To check the most powerful AI systems, their testers will also themselves have to be powerful AI systems, precisely trained and refined to excel at the single task of identifying safety concerns and problem areas in the world’s most advanced models.

To be trustworthy and yet agile, these testing companies should be checked and certified by government regulators but developed and funded in the private market, with possible support by philanthropy organizations

The field is moving too quickly and the stakes are too high for exclusive reliance on typical government processes and timeframes.

One way this can unfold is for government regulators to require AI models exceeding a certain level of capability to be evaluated by government-certified private testing companies (from startups to university labs to nonprofit research organizations), with model builders paying for this testing and certification so as to meet safety requirements.

As AI models proliferate, growing demand for testing would create a big enough market. Testing companies could specialize in certifying submitted models across different safety regimes, such as the ability to self-proliferate, create new bio or cyber weapons, or manipulate or deceive their human creators

Much ink has been spilled over presumed threats of AI. Advanced AI systems could end up misaligned with human values and interests, able to cause chaos and catastrophe either deliberately or (often) despite efforts to make them safe. And as they advance, the threats we face today will only expand as new systems learn to self-improve, collaborate and potentially resist human oversight.

If we can bring about an ecosystem of nimble, sophisticated, independent testing companies who continuously develop and improve their skill evaluating AI testing, we can help bring about a future in which society benefits from the incredible power of AI tools while maintaining meaningful safeguards against destructive outcomes.

“What Putin wants is to transform the world order” that evolved since World War II and the post-Cold War — where “the competition between nations was about who can be richer and who can help their people prosper the most . … Putin hates that world because he loses in that world — his system is a loser in a peaceful, global, wealth-enhancing paradigm. And so what he wants is to move us back to dog-eat-dog, to a 19th-century, great power competition, because he thinks he can, if not win, be more effective there. … Let’s not think that this is a Ukrainian problem; this is a problem for us all.”

These wars very much are our business — and now clearly inescapable, since we’re deeply entwined in both conflicts. What’s crucial to keep in mind about America — as the leader of the Inclusion Network — is that right now we’re fighting the war in Ukraine on our terms, but we’re fighting the war in the Middle East on Iran’s terms.

CNN recently described, per a source familiar with it, a declassified U.S. intelligence assessment provided to Congress saying that Russia had lost 87 percent of its preinvasion active-duty ground troops and two-thirds of its tanks that it had prior to its invasion of Ukraine. Putin can still inflict a lot of damage on Ukraine with missiles, but his dream of occupying the whole country and using it as a launching pad to threaten the Inclusion Network — particularly the NATO-protected European Union — is now out of reach. Thank you, Kyiv.

At a breakfast with NATO leaders devoted to the Ukraine issue at Davos this year, Canada’s deputy prime minister, Chrystia Freeland, noted that it is we, the West, who should be thanking the Ukrainians, not forcing them to beg us for more weapons.

China under President Xi Jinping straddles the two networks, along with much of what’s come to be called the global south. Their hearts, and often pocketbooks, are with the Resistors but their heads are with the Includers

the Resistance Network “is orchestrated by Iran, Islamists and jihadists” in a process they refer to as the “unity of battlefields.” This network, he noted, “seeks to bridge militias, rejectionists, religious sects and sectarian leaders,” creating an anti-Israel, anti-American, anti-Western axis that can simultaneously pressure Israel in Gaza, in the West Bank and on the Lebanon border — as well as America in the Red Sea, in Syria and in Iraq and Saudi Arabia from all directions.

In stark contrast, Koteich said, stands the Inclusion Network, one that’s focused on “weaving together” global and regional markets instead of battlefronts, business conferences, news organizations, elites, hedge funds, tech incubators and major trade routes. This inclusion network, he added, “transcends traditional boundaries, creating a web of economic and technological interdependence that has the potential to redefine power structures and create new paradigms of regional stability.”

things are different in the Middle East. There, it is Iran that is sitting back comfortably — indirectly at war with Israel and America, and sometimes Saudi Arabia, by fighting through Tehran’s proxies: Hamas in Gaza, the Houthis in Yemen, Hezbollah in Lebanon and Syria, and Shiite militias in Iraq.

Iran is reaping all the benefits and paying virtually no cost for the work of its proxies, and the U.S., Israel and their tacit Arab allies have not yet manifested the will or the way to pressure Iran back — without getting into a hot war, which they all want to avoid.

The members of the Resistance Network are great at tearing down and breaking stuff, but, unlike the Inclusion Network, they have shown no capacity to build any government or society to which anyone would want to emigrate, let alone emulate

For all of these reasons, this is a moment of great peril as well as great opportunity — especially for Israel. The competition between the Resistance Network and the Inclusion Network means that the region has never been more hostile or more hospitable to accepting a Jewish state.