Dr. Ben Goertzel on ai: what the evidence says · JRE #1211

What the evidence says 01 / RECORD

In this 2019 interview, Ben Goertzel predicted human-level artificial general intelligence (AGI) would be achieved within five to seven years, placing the target at roughly 2024-2026. As of mid-2026, there is no consensus that human-level AGI has been achieved: Scientific American reports that despite AI systems clearing specific benchmarks, such as a large language model reportedly passing a Turing test in 2025, complex reasoning remains a persistent gap and researchers keep redrawing the goalposts on what counts as 'real' intelligence. Nature similarly reports that current models such as OpenAI's o1 show more sophisticated reasoning but are unlikely to reach human-level general intelligence through scaling alone. A minority of industry figures have asserted that forms of AGI already exist, or argued the concept no longer resists a single moment of arrival, but this is a contested claim amid ongoing disagreement among researchers over how AGI should even be defined, not an established scientific finding. Goertzel's own AGI research (through SingularityNET and OpenCog) did not deliver human-level AGI within his stated 2024-2026 window; in a June 2026 Forbes interview he said he now expects human-level AGI around 2029, with 2027-2030 as his uncertainty range, roughly matching Ray Kurzweil's long-standing 2029 estimate. The specific five-to-seven-year timeline given in this claim has not been borne out.