Warshaw warns: AI could flood the market with mediocre games

Nov 21, 2026 5 min

Howard Scott Warshaw compares the flood of AI-created games to the glut before the 1983 crash, though he dismisses the idea that the industry will disappear.

Howard Scott Warshaw, creator of E.T. for Atari 2600, believes artificial intelligence could multiply the number of mediocre games until it recalls the glut that preceded the 1983 crash. His warning, delivered at DevGAMM Lisbon, comes with a decisive caveat: he does not expect today's industry to collapse as it did then, although he does believe the abundance will change how games are discovered and valued.

The comparison is not a prediction of another financial crisis, but a warning about the relationship between more games and more noise. If production becomes faster, releases of uneven quality may also increase; for Warshaw, the risk is that worthwhile projects will be buried by the sheer volume.

A warning born of Atari

Warshaw gave the talk “From Atari to AI” at DevGAMM, the video game development conference held in Lisbon on November 17 and 18. The developer is also known for Yar’s Revenge, but his connection to E.T. makes his perspective on the crash particularly direct: the game is cited among the factors associated with the 1983 crisis.

The scale of that collapse was enormous: industry revenues fell from more than $3 billion in 1983 to around $100 million in 1985, according to Hobby Consolas. Warshaw links that episode to a market open to too many low-quality releases, a situation that, in his view, could find an echo in AI-assisted production.

“will have the same impact, but on a larger scale”

— Howard Scott Warshaw, creator of E.T. for Atari 2600

The remark points to the scale of the problem, not a literal repeat of history. In later comments reported by Hobby Consolas, Warshaw said he sees no prospect of an equivalent crisis: in the early 1980s, people questioned whether video games were anything more than a fad, whereas today they are an established part of everyday life.

AI speeds up production; it does not guarantee the result

The change that concerns him stems from a specific advantage. Warshaw acknowledged that AI can significantly boost programming: tasks that once took weeks, months, or years can be completed in days with its help. That time savings opens the door to producing more games, and that is where he sees a parallel with the expansion in releases that preceded the Atari crisis.

But speed and quality are separate matters. App2Top summarized Warshaw's position as ambivalent: these technologies simplify and accelerate development, but they also affect the quality of projects. The concern, then, is not that the tool will make every game bad, but that it will reduce the effort needed to bring more products to market without determining on its own which ones deserve attention.

This leaves studios and companies facing a practical tension: AI can shorten programming stages, while a more abundant supply can make it harder for each release to find an audience. His prediction is broader: companies will have to find effective ways to distinguish good games from bad ones.

Why 1983 is not a blueprint

The key difference from the past is confidence in the medium's continued existence. Warshaw explained that during the crisis of the 1980s, some investors stopped backing development because they did not know whether video games would last. Today, according to his argument, that doubt about the industry's survival no longer carries the same weight: video games are firmly established and present in people's lives.

That is why his concern is not equivalent to saying that AI will erase the industry. Warshaw considered that video games might become “something completely different,” but rejected the idea that they will disappear. App2Top described his expectation as a restructuring: the market will have to learn to filter and bring standout projects to the fore.

The distinction matters when interpreting the analogy. Warshaw sees a similarity in the glut of low-quality games, but does not claim that the economic conditions of 1983 will recur. The earlier crash combined oversaturation with doubts about whether the medium had a future; his argument is that the second factor no longer operates in the same way.

The challenge shifts from creating to standing out

For developers, the opportunity Warshaw describes is clear, though limited: AI assistance can speed up programming and make it possible to reach results sooner than before, when they required much more time. His warning, however, focuses on the collective cost of that convenience: each team can produce more, but the market can also fill up with more releases competing for attention.

For companies, the issue is no longer just how long it takes a team to build a game. If supply increases, project selection and the ability to help quality games avoid being buried become more important. App2Top summarizes Warshaw's proposal as the need for companies to find effective ways to distinguish good games from bad ones.

The conversation at DevGAMM also took place at an industry-focused event, featuring sessions, a game showcase, and business meetings. The organizers announced that the 2026 edition would bring together 85 speakers and showcase more than 100 games. In that context, Warshaw's warning is not a call to stop AI: it asks what happens when producing more is no longer enough of an advantage to help a game find its place.

His comparison with Atari points to pressure on the market, not a date for a new collapse. The open question is whether companies will find ways to separate projects that deserve attention from the flood that, according to Warshaw, AI could accelerate.

Sources: Hobby Consolas, App2Top, DevGAMM Lisbon 2026 - International Gamedev Conference