A recent decision by HackerOne to suspend new vulnerability submissions to its crowdsourced Internet Bug Bounty (IBB) program has spotlighted the growing remediation challenges across the industry, driven by the rapid advancement of AI-assisted bug hunting and discovery.
Launched in 2013, the IBB is widely regarded as one of the open source community’s most important vulnerability reward programs. Effective March 27, the program paused accepting new vulnerability submissions because of what HackerOne described as a worsening imbalance between vulnerability discoveries and the ability for open source maintainers to remediate them.
‘Signal Versus Noise’
“The discovery landscape is changing. AI-assisted research is expanding vulnerability discovery across the ecosystem, increasing both coverage and speed,” HackerOne announced. “The balance between findings and remediation capacity in open source has substantively shifted,” and requires a rethink of the structure and incentives of crowdsourced programs like IBB, it said.
Following HackerOne’s decision, the maintainers of the open source Node.js project paused their own bug bounty program, citing a loss of funding previously available via HackerOne. “As a volunteer-driven open-source project, Node.js does not have an independent budget to sustain a bounty program on its own,” the maintainers explained.
Several security experts perceived the announcements as significant but not unexpected, given the rapid growth of AI-assisted vulnerability discovery.
“This is a rational, even overdue correction to how bug bounty ecosystems operate under AI pressure,” says Ensar Seker, chief information security officer (CISO) at SOCRadar. “HackerOne is essentially acknowledging that the bottleneck has shifted: discovery has been industrialized by AI, but remediation capacity has not scaled accordingly,” he says. When AI can generate thousands of low- to medium-quality findings in a matter of hours, the maintainers of open source projects, who are often volunteers with limited funding, can easily get overwhelmed. “So yes, [HackerOne] absolutely have a point that this is not a retreat from security, it’s an attempt to rebalance signal versus noise.”
AI-Generated Slop
John Morello, co-founder and chief technology officer (CTO) of Minimus, says valid submissions dropped from roughly 15%, to below 5%, as AI-generated “slop” floods the gates. “AI-assisted hunting hasn’t necessarily found more critical zero-days; instead, it’s shifted the bottleneck entirely to validation, forcing triage teams to wade through thousands of plausible-sounding but non-exploitable reports,” he says.
For open source project maintainers, “triage fatigue” has become the biggest challenge where they are losing hours of development time just to disprove hallucinated vulnerabilities. “The current bounty model unfortunately rewards quantity over depth, effectively weaponizing unpaid labor and forcing these small teams to act as a free [quality assurance] department for every automated scanner on the planet,” Morello says.
HackerOne said its focus now is on finding new ways to meet its original objectives of aligning vulnerability discovery with effective remediation “so that meaningful findings lead to durable security improvements in open source projects.” To that end, HackerOne will work with project maintainers and researchers to evaluate approaches that better align incentives with the realities of the open source ecosystem.
Trey Ford, chief strategy and trust officer at Bugcrowd, which also operates a crowdsourced vulnerability discovery platform, perceives HackerOne’s decision as a wakeup call. “Let’s be clear about what this pause actually signals: the industry spent years optimizing the wrong end of the pipeline,” he says. AI has done exactly what it was supposed to do in terms of compressing the time required to find vulnerabilities. “What we have not yet solved is the human side of the equation: the maintainer who receives 40 valid reports and has one weekend to respond,” Ford says.
Discovery’s Funded, Remediation Isn’t
What’s required now are greater investments in remediation capacity with the same urgency as has been put into discovery. “The economics of research and disclosure are shifting. AI lowers the barrier to finding, which means raw volume is no longer a competitive advantage for researchers,” Ford notes. The premium increasingly will move toward complex logic flaws and novel attack chains that require human depth and contextual judgment that machines cannot replicate. “The next generation of vulnerability programs may offer bonuses to researchers for bringing fixes, not just reporting vulnerabilities, and create shared pools that fund both the researcher who finds and the maintainer team that ships the patch.”
Remediation is not the only challenge. As David Hayes, VP of product at FusionAuth notes, bug bounty programs designed around human-paced research are burning through funds faster than anyone anticipated. “The model as currently structured isn’t sustainable,” he says. Bounties were designed for a world where discovery was the bottleneck. Now that discovery is increasingly automated, the bottleneck is remediation, which bounties don’t fund. “The projects that underpin critical Internet infrastructure can’t rely on volunteer labor to process AI-generated reports at scale,” he says. “The industry needs to figure out how to fund the fix, not just the find.”
