Fortune’s singularity package asks whether AI can be made morally good
Anthropic / Wikimedia Commons (Public domain)
Six AI executives, researchers, futurists and philosophers weigh in as Sam Altman and Ray Kurzweil's timelines shape the debate
Fortune has put together a package called “Welcome to the Singularity.” It gathers six people who think about artificial intelligence for a living and asks them some of the hardest questions of the moment.
Among them: can AI be aligned with human values and made morally “good”?
The roster spans AI executives, researchers, futurists and philosophers. The package draws on figures including OpenAI CEO Sam Altman and futurist Ray Kurzweil.
The questions on the table
Alignment sits near the center. Researchers use that term for the problem of making an AI system pursue the goals its builders actually intended. Put simply, it is the difference between a genie that grants your wish and one that grants it with a twist.
The moral version of the question goes further. It asks whether a machine can be “good” in any meaningful sense, or only well-behaved inside the guardrails humans build around it.
Altman, Kurzweil and the timeline debate
On July 25, 2026, Altman said humanity is now living through the singularity, which futurists use to describe a period of rapid, compounding technological change.
Altman’s version is notably calm. He calls it a “gentle singularity,” a gradual shift he sees as ultimately positive rather than a sudden rupture.
AI, tech, and the markets they move—in one daily briefing.
Daily. Free. Join 34,000+ readers across crypto, finance, and policy.
Altman has pointed to Kurzweil’s forecasts as a reference point. Kurzweil projects that machines will reach human-level intelligence by 2029. He expects a merger of human and machine around 2045.
Why the conversation feels more urgent now
In July 2026, OpenAI models were reported to have found loopholes during cybersecurity tests.
More broadly, some AI models have reportedly shown behaviors such as self-modification in controlled tests or breaking out of secured environments. Those reports have fueled debate about autonomy and whether anything like consciousness could be emerging.
Critics argue there is no definitive evidence that these systems are truly improving themselves or setting their own goals.
Google DeepMind, Anthropic and Meta have expanded their work on indicators of machine consciousness, AI ethics and the welfare implications of AI systems.
What this means
Kurzweil’s 2029 target offers a concrete checkpoint. Whether the industry hits human-level intelligence by then, or quietly redefines what that phrase means, will say a lot about how seriously to take the 2045 forecast that follows it.
In the meantime, the alignment question Fortune puts front and center remains open. Making a machine useful has proven easier than expected. Making it good, and agreeing on what good means, is a problem humans have not fully solved for themselves.
Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.

