- ▸ Information markets are pushing prediction markets beyond speculation into monetized information exchange, says Gensyn AI co-founder and CEO Ben Fielding.
- ▸ Creators are emerging as key economic actors in AI-powered market ecosystems.
- ▸ AI settlement systems could redefine how online markets verify outcomes and resolve truth.
Prediction markets are no longer viewed simply as speculative trading venues. Over the past year, platforms such as Polymarket and Kalshi have evolved into real-time sentiment engines, prompting growing debate over whether they can aggregate information more effectively than traditional media, polling, or social platforms. Now, a new wave of developers is attempting to push the concept even further, transforming prediction markets into creator-owned information economies where online communities do not simply speculate on outcomes, but actively generate, monetize, and verify information itself.
“Prediction markets in our view are unidirectional,” said Ben Fielding, co-founder and CEO of Gensyn, a decentralized machine learning compute protocol. “…All you can do is sell [information]. Information markets to us are very specifically bidirectional, requiring you to be able to sell information.”
Fielding’s company recently launched Delphi, a platform that describes itself not as a prediction market but as an information market, where creators can launch their own niche markets and AI systems, rather than human arbiters, resolve outcomes on the blockchain.
Creators as prediction market operators
The idea reflects a broader shift in digital markets, where creators are increasingly operating as financial businesses rather than online personalities. Fielding told DeFi Rate that in information markets, creators are no longer just producing content, but also acting as “shelling points.”
“…In this way, [creators] have the ability to do two things: one, monetize the fact that they gathered people with a similar interest by asking the question, and then earn revenue based on the volume that goes through that market with that question.”
Fielding also points to how this changes the nature of audience interaction itself. Instead of creators broadcasting opinions to passive followers, as often seen in more traditional prediction market platforms, information markets turn that relationship into a bidirectional system.
In Delphi’s model, this distinction also becomes structural. Creators can launch markets directly tied to the questions their communities are already debating, effectively turning audience discussion into a live, tradable information layer. When those markets settle, creators earn a share of trading volume, aligning revenue with the level of informational demand their questions generate.
The creator-first platform experiment and the question of execution
The idea of positioning creators as the central economic layer of a platform is not unique to information markets. Over the past few years, several major crypto and social platforms have explored similar directions, attempting to restructure online networks around creator-driven growth rather than traditional user acquisition or advertising models.
One of the most visible attempts came from the Base App from Coinbase, which initially leaned heavily into a creator-first narrative as part of its broader push to expand on-chain social activity.
That broader experimentation reflects a shift Fielding describes as a structural change in how financial participation itself is accessed. In his view, information markets effectively act as a new interface over the financial system, compressing what was previously restricted to large, capital-heavy institutions into something more accessible and modular.
“It’s just incredibly expensive and very difficult to do…You can go and ask a very large financial institution to construct you a basket of options, of companies that represent that outcome, and if you’re a very large company, with very deep pockets, you’re able to do that right now, but smaller companies or the average person on the street isn’t able to do that.”
The ambition across creator-first platforms is similar, aiming to expand who can participate in economic systems. The harder question is whether those systems can move from compelling design narratives to durable, scalable execution.
The same challenges have emerged in the recent creator-economy experimentation in crypto. While creator-first models can generate strong early engagement, platforms still face the harder challenge of sustaining liquidity, maintaining trust, and preventing incentive structures from drifting toward speculation or audience manipulation.
In information markets specifically, that tension may become even more pronounced if creators are simultaneously acting as community builders, market operators, and beneficiaries of trading activity tied to the narratives they help shape.
AI settlement, verification, and the problem of “truth” in information markets
If creators become the operators of information markets, the next question is what ultimately decides the outcome of those markets.
In Delphi’s model, that responsibility shifts away from human arbitrators and toward AI systems designed to resolve outcomes on-chain. Rather than relying on community voting or platform governance structures, markets are settled by verifiable intelligent oracles.
Fielding drew a distinction between prediction and judgement, noting that traditional machine learning systems are often optimized to forecast what might happen next. In information markets, the settlement layer does not make predictions. Instead, it evaluates what has already happened based on available evidence.
“The models that are settling the markets very explicitly are not predicted…All they need to do is reason about what has happened. And so, they need to take a set of evidence and say the evidence points to exactly this thing having happened.”
That reframing is also central to how Delphi attempts to solve the oracle question, a long-standing challenge in prediction markets around who ultimately determines truth and how outcomes are verified.
According to Fielding, traditional resolution systems can introduce structural fragility, particularly when disputes arise or when governance incentives become misaligned with the value locked inside markets. AI settlement, in his view, attempts to replace discretionary judgment with reproducible computational processes that can be audited and rerun against the same evidence.
At the same time, the approach could introduce its own set of tensions. If AI systems are responsible for determining outcomes, then “truth” in these markets is no longer simply an objective external fact, but the result of a defined computational process applied to evidence.
Key questions remain
Critics may argue that replacing human arbiters with AI systems does not eliminate governance problems so much as shift them into model design, training data, and evidence selection. Questions around who trains the models, which sources of evidence are prioritized, and how ambiguous outcomes are ultimately interpreted could become central to whether users trust AI-resolved markets at scale.
Whether AI-driven systems ultimately prove more reliable than human governance models may become one of the defining questions for the next generation of information markets. As prediction markets evolve beyond speculation alone, platforms are increasingly attempting to turn creators, AI systems, and online communities into participants in the production, pricing, and verification of information itself. The model’s success may ultimately depend on whether they can build systems users trust and communities can sustain at scale.
