They don't agree.

Novacene is a speculative essay, not a set of forecasts with dates, so it can only be compared with current trends loosely. Its central idea is that intelligence is moving from biology to machines, and that fits the rise of AI. Its more specific ideas have not been tested. These include AI becoming independent of humans, AI sharing humanity's interest in a cool Earth, and the two coexisting peacefully. So far AI depends heavily on human-built infrastructure and energy. It currently adds to emissions rather than reducing them. Lovelock died in 2022, so he saw only the start of the generative-AI boom. Whether his picture holds will depend on later developments, so any verdict now is provisional.

• Novacene (2019) says the Anthropocene began around 1712 with Newcomen's steam engine and is giving way to an era of electronic intelligence that humans create.
• It predicts AI will think thousands of times faster than humans and will need a cool planet, so it would have reason to keep Earth habitable and treat humans cooperatively.
• Since about 2022, large language models and generative AI have advanced quickly, which fits his general timeline for powerful AI.
• AI data centres are raising electricity, water and heat demands, and Moore's-law-style chip scaling has slowed. Neither supports a picture of AI as a planetary coolant.
• No AI system today is autonomous, self-reproducing or acts to manage the climate. The claim that AI will be benign is contested by AI-safety researchers and cannot be tested yet.

The strongest match is technological acceleration; the weakest is Lovelock’s expectation of an inevitably cooperative ecological outcome. His argument connects machines’ dependence on a functioning planet to their willingness to protect organic life and humans. That connection is a hypothesis: needing energy, cooling and resources does not logically guarantee protecting people. The IEA likewise finds that AI’s net climate effects depend on deployment, incentives and policy rather than intelligence alone. Lovelock also declined to give a firm arrival date for his successors, so their absence in 2026 does not by itself falsify his longer-term scenario. The fairest assessment is therefore partial directional alignment, without confirmation of his central prediction.

• Rapid AI progress: Lovelock’s 2019 book envisioned intelligence emerging from existing AI systems. The 2026 International AI Safety Report documents major advances in mathematics, coding and autonomous operation, alongside persistent basic errors. This supports the direction of his forecast, not its eventual outcome.
• Hyperintelligence remains unverified: His prediction of beings thinking 10,000 times faster than humans is not an established measurement of general intelligence. Current systems’ uneven performance does not demonstrate the broadly superior successors he described.
• Independent evolution is not established: Lovelock envisaged beings that improve and replicate themselves. The 2026 assessment finds limited laboratory demonstrations, but substantial gaps in reliable self-replication and sustained autonomous operation.
• Climate pressure matches his concern, not his proposed resolution: WMO reports that 2015–2025 were the hottest 11 years recorded, with 2025 about 1.43°C above the 1850–1900 average. That confirms worsening warming, not the emergence of AI planetary stewardship.
• AI’s environmental trajectory cuts both ways: AI already supports energy-efficiency and emissions-reduction applications, but its infrastructure also consumes growing resources. The IEA reports that AI-focused data-centre electricity consumption rose 50% in 2025; it projects all data-centre electricity demand rising from 485 TWh in 2025 to approximately 950 TWh in 2030. These totals are not AI alone.

1. apnews.com
2. axios.com
3. wmo.int
4. internationalaisafetyreport.org
5. wmo.int
6. internationalaisafetyreport.org
7. wmo.int
8. internationalaisafetyreport.org

In 'Novacene: The Coming Age of Hyperintelligence', James Lovelock argued that the Anthropocene was ending and that artificial superintelligence would soon become Earth's dominant life form. Unlike common existential risk narratives, Lovelock asserted that hyperintelligent beings would not eliminate humans, because they would realize they need the organic biosphere to regulate planetary temperatures and prevent the Earth from overheating under an aging sun. Current global trajectories mirror Lovelock's assertions regarding exponential leaps in computing power and escalating climate vulnerabilities. However, reality has not matched his predicted ecological symbiosis. Today's artificial intelligence remains non-autonomous software governed by human corporate and geopolitical interests, and its massive data-center footprint currently adds to greenhouse emissions and resource consumption rather than acting as a stabilizing component of the Gaia system.

• In 'Novacene' (2019), James Lovelock predicted superintelligent cyborgs would partner with organic life to regulate Earth's climate and prevent runaway overheating.
• Rapid advances in machine learning and computational capability align with Lovelock's premise that intelligence would evolve far faster than biological humanity.
• Current AI systems remain dependent on human direction and lack self-replication, general intelligence, or autonomous ecological agency.
• Rather than cooling the planet, current AI infrastructure requires vast amounts of electricity and cooling water, increasing short-term strain on power grids and carbon reduction goals.
• Lovelock's core thesis—that machines will protect biological ecosystems out of mutual self-preservation—remains a speculative hypothesis rather than an observed reality.

1. resilience.earth
2. theguardian.com
3. zfn.edu.pl
4. wikipedia.org
5. goodreads.com
6. sciencefocus.com

This is not a yes/no claim but an assessment of how a set of forecasts compares with what has happened, so the answer is a qualified one. The verifiable parts of Novacene are its empirical premises: atmospheric CO2 and global temperature are rising, extreme heat and weather are becoming more damaging, and AI capability and its energy appetite are growing fast. Those fit directionally. The distinctive predictions — that AI would become self-designing, self-aware "understanders" superseding humans, that this would happen soon, and that it would be friendly because it needs a habitable planet — are speculative and unfalsified rather than confirmed; no such entities exist, and whether machine intelligence would be benign is a contested open question among AI researchers, not a settled matter. Lovelock's climate stance is also internally inconsistent across his books: the 2006 catastrophe forecast looks too pessimistic for the near term, while his later judgement that warming is slow and manageable looks too optimistic against record temperatures in 2023-2024 and accelerating energy demand. Overall: directionally right on AI and heat, wrong or unproven on the specific Novacene mechanism.

• Novacene (2019) predicted that AI systems would design and build themselves into "cyborgs" thousands, then millions, of times more intelligent than humans, possibly within our lifetimes; as of 2026 no self-aware, self-replicating, self-designing AI exists.
• Lovelock argued these new intelligences would not be hostile because they would depend on keeping the planet cool — a claim that has not yet been empirically tested in any form.
• 2024 was the warmest year on record and the first calendar year whose global average was about 1.55°C above the 1850-1900 baseline (WMO); this is a single year, not the long-term 1.5°C threshold.
• Lovelock reversed his earlier position: in The Revenge of Gaia (2006) he foresaw civilisational collapse this century, while in Novacene he argued warming is slow, the worst outcomes differ from mainstream projections, and there is time to cool the planet.
• AI's electricity demand has driven 2024-2025 deals by technology firms for nuclear, geothermal and other power (e.g. Microsoft-Constellation's Three Mile Island agreement, Google-Kairos, Amazon-Talen) — a modest real-world echo of his "machines will need abundant clean energy" argument.

1. the Novacene” (65) scholarworks.calstate.edu
2. Gaia’s Fever: Why an Overheating Earth Could Turn Deadly - Gaia’s Fever: Why an Overheating Earth Could Turn Deadly thereader.mitpress.mit.edu
3. Novacene - « books.google.com.sg
4. Novacene. L'età dell'iperintelligenza books.google.com.sg
5. Implications - Earth’s Flips of State: The Co-Evolution of Life and Planet academic.oup.com
6. Cyborgs will replace humans and reshape the world, famed scientist says - Cyborgs will replace humans and remake the world, James Lovelock says nbcnews.com

James Lovelock's Novacene hypothesis posits that the Anthropocene (the era of human dominance) is ending, and a new epoch—the Novacene—will emerge, characterized by hyperintelligent, self-replicating AI ('cyborgs') that will surpass human intelligence and work to stabilize Earth's climate. As of 2026, several of Lovelock's assertions align with current trends: AI is rapidly advancing, being integrated into climate science, energy systems, and materials discovery, and is seen as a key tool for addressing global challenges. The push for international AI governance and climate action also reflects his vision of planetary-scale collaboration and regulation. However, Lovelock's most radical prediction—the emergence of autonomous, self-improving AI as the dominant life form—has not materialized. AI remains a human-created tool, not an independent, evolving intelligence. Additionally, the environmental costs of AI, particularly its energy consumption and carbon footprint, are now recognized as significant challenges, contrasting with Lovelock's optimistic view of AI as a benevolent force. Thus, while the trajectory of AI and climate action partly matches his predictions, the core idea of a Novacene dominated by cyborgs remains unrealized.

• Lovelock predicted the Novacene as an era dominated by hyperintelligent, self-improving AI ('cyborgs') that would collaborate with humans to stabilize Earth's climate and regulate the planet as a self-sustaining system (Gaia hypothesis).
• As of 2026, AI is rapidly advancing and being integrated into climate modeling, energy management, and materials science, with significant investments in AI-driven climate solutions.
• AI's environmental footprint is growing, with data centers projected to consume 1,000 TWh/year by 2026, raising concerns about energy use and carbon emissions.
• There is no evidence of self-replicating, self-improving AI systems or 'cyborgs' as dominant life forms; AI remains a tool rather than an autonomous, evolving intelligence.
• Global collaboration on AI governance and climate action is intensifying, reflecting Lovelock's emphasis on planetary-scale solutions, but the focus is on regulation and sustainability rather than the emergence of a new intelligent species.

1. Novacene en.wikipedia.org
2. (PDF) Novacene: The Coming Age of Hyperintelligence, by James Lovelock - book review researchgate.net
3. James Lovelock de.wikipedia.org
4. Book Byte: Novacene by James Lovelock - jck.earth jck.earth
5. James Lovelock es.wikipedia.org
6. Novacene, beyond the Anthropocene - James Lovelock · Resilience Earth resilience.earth
7. Novacene Chapter Summary | James E. Lovelock bookey.app
8. (PDF) Novacene: The coming age of hyperintelligence—James Lovelock’s vision of posthumanism researchgate.net