The threat of deepfakes is growing at an alarming rate, with a 245% increase in the number of deepfakes worldwide between 2023 and 2024. This surge in deepfakes has been spurred in part by upcoming election cycles, according to verification provider Sumsub.
Deepfakes: A Growing Concern for Businesses
The trend is not limited to politics; deepfakes are also affecting the corporate sector. A recent survey conducted by Business.com found that 10% of companies have faced deepfake-aided fraud, such as cloned voices. This has led to a significant increase in demand for companies marketing tools to defend against deepfakes and technologies used to produce them.
Pindrop Secures $100 Million Loan from Hercules Capital
One company that is well-positioned to take advantage of this trend is Pindrop, which recently announced that it had secured a $100 million, five-year loan from Hercules Capital. The funds will be used for product development and hiring, according to CEO Vijay Balasubramaniyan.
The Need for Deepfake Detection
Balasubramaniyan noted that advancements in generative AI have made voice cloning a powerful tool, and deepfake detection leveraging AI detection technologies is now needed by every call center to stay one step ahead of fraudsters. Pindrop builds deepfake-combating and multi-factor authentication products targeting businesses in banking, finance, and related industries.
How Pindrop’s Tools Work
Pindrop’s tools can identify contact center callers with higher accuracy than rival solutions. The company leverages a dataset of over 20 million utterances, both synthetic and genuine, to train the AI models to differentiate between genuine human voices and synthetically generated ones.
The Challenge of Bias in Deepfake Detection
Bias is a common problem in deepfake detection models. Many audio models tend to be biased towards recognizing Western, American voices and perform poorly with different accents and dialects, which could lead a detector to classify a legitimate voice as a deepfake. It’s up for debate whether synthetic training data – training data generated by AI models themselves – mitigates or exacerbates the biases.
Pindrop’s Approach to Addressing Bias
Balasubramaniyan thinks that synthetic training data can actually help mitigate bias, and Pindrop’s voice authentication products focus on acoustic and spectro-temporal features of voices as opposed to pronunciation or language. However, he acknowledges that AI-based voice recognition systems tend to display biased results towards tonal, accentual, and dialectical variations.
The Importance of Addressing Bias
Addressing bias is crucial in deepfake detection, as it can have serious consequences for individuals and businesses. For example, a deepfake video or audio clip could be used to manipulate public opinion or deceive investors.
Pindrop’s Future Plans
With the $100 million loan from Hercules Capital, Pindrop plans to continue developing its deepfake-combating technologies and expanding its product offerings. The company aims to become a leading provider of deepfake detection solutions for businesses around the world.
Conclusion
The threat of deepfakes is growing at an alarming rate, but companies like Pindrop are working tirelessly to develop countermeasures. With advancements in generative AI making voice cloning more powerful than ever, it’s essential that businesses invest in deepfake detection technologies to protect themselves from this emerging threat.
Related News
- Venture: What will this year bring in VC? We asked a few investors
- Startups: Inside the wild fall and last-minute revival of Bench, the VC-backed accounting startup that imploded over the holidays
- Fintech: These fintech companies are hiring in 2025 after a turbulent year
Subscribe to TechCrunch
Stay up-to-date with the latest news and trends in the tech industry by subscribing to TechCrunch’s newsletters. You can choose from a range of topics, including startups, AI, space, and more.
