Password-Less with AI
Last Update: Sep 23, 2023
Here's What I Found in My Gmail Today!
The imminent arrival of the next generation of mainstream AI applications is upon us. The initial wave introduced us to popular applications such as ChatGPT, which focused on text generation, and tools like Midjourney, which specialized in image generation.
However, this upcoming wave promises to be truly multimodal, featuring applications capable of both receiving inputs and generating outputs in various formats, including text, images, videos, and more.
There have been speculations circulating that Google might beat OpenAI to the multimodal arena with their new Gemini model. Still, OpenAI has taken the lead by introducing Dall-E 3, a cutting-edge AI image generation model designed to seamlessly integrate with ChatGPT for image creation.
These technical advancements are not just about showcasing innovation; they also bring tangible benefits to users like us:
Improved Output Quality:
OpenAI claims that Dall-E 3 can comprehend prompts more profoundly, resulting in the creation of highly detailed images. Some early-access users on Twitter seem to concur with this claim.
Simplified Prompt Generation:
Unlike other mainstream AI apps, such as Midjourney, which often demand expertise in crafting prompts, Dall-E 3 is integrated directly into ChatGPT. You can simply instruct ChatGPT regarding what you want to visualize, and it will generate prompts for you.
Text Integration in Images:
Unlike other AI image models that struggle with accurately incorporating text into images, Dall-E 3 asserts its ability to do so effectively. This means users can produce items like brand logos, graphic advertisements with text for marketing campaigns, and other assets combining text and images for professional and personal use.
OpenAI has stated that images generated using Dall-E will be available for free use in commercial applications, eliminating the need for permissions or legal concerns. This allows users to incorporate these images into their work without worry.
Dall-E 3 is scheduled to roll out to ChatGPT Plus and Enterprise users in a few weeks. Additionally, OpenAI has pledged to give artists the option to exclude their art from future iterations of text-to-image AI models.
In the realm of AI and security, there is a growing trend of using artificial intelligence to enhance the sophistication and targeting of various cyberattacks, including phishing, account takeovers, and credential stuffing. This shift underscores the critical need for businesses to prioritize security measures for both themselves and their users.
As AI capabilities continue to evolve, new threats are emerging in online security, affecting consumers in several ways:
AI models, like PassGAN, can crack a significant percentage of common passwords rapidly, highlighting the need for stronger authentication methods.
AI tools can now steal passwords with high accuracy by eavesdropping on keystrokes, posing a risk in public spaces or during phone calls.
Large language models have empowered phishing attacks, making it harder to distinguish genuine communications from fraudulent ones.
To address these challenges, businesses must adopt robust security practices:
- Consumer Security Education: Apps should educate users about security risks and offer reporting mechanisms for suspicious activity.
- Multi-Factor Authentication (MFA): Implementing MFA adds an extra layer of security, but it should be balanced with user convenience to avoid churn.
- Account Takeover Recovery Support: Businesses should provide clear guidance on recovering accounts and handling losses resulting from attacks.
- Passkey Authentication: Passkeys, based on FIDO Alliance standards and supported by major platforms, offer a secure and user-friendly alternative to traditional passwords.
Passkeys enhance security by utilizing public-key cryptography and biometric authentication, offering several advantages such as resistance to replay attacks, site-specific authentication, and no storage of sensitive credentials on servers.
Passkey adoption is on the rise, with major platforms and websites like Google, Apple, LinkedIn, and others either implementing or considering passkey support for billions of users. However, implementing passkeys can be challenging, leading many organizations to turn to services like Passage by 1Password for streamlined implementation and maintenance.
Passage by 1Password simplifies passkey support and offloads the complexity to authentication experts, allowing businesses to enhance security without sacrificing the user experience. It's a transparent solution that benefits both 1Password customers and non-customers alike, making passkeys accessible to a broader audience with minimal effort.
The next wave of mainstream AI apps is upon us. The first wave brought us popular apps like ChatGPT which generated text and tools like Midjourney which generated images. This next wave is going to be multimodal: apps that can both take inputs and produce outputs in various formats like text, images, video and so on.
And while rumors have been circulating that Google may be able to beat OpenAI to the multimodal party with their new Gemini model, OpenAI has landed the first punch with Dall-E 3: a brand new AI image generation model that will integrate with ChatGPT to produce images.
Technical advances are cool. But here’s what it’s going to help users like us:
OpenAI says that Dall-E 3 is able to understand prompts more deeply and produce higher detail images. Some Twitter users with early access agree.
While other mainstream AI apps like Midjourney often require expert prompting knowledge, Dall-E 3 is built natively on ChatGPT. Just tell ChatGPT what you want to see, and it will generate prompts for you.
Text in Image:
While other AI image models struggle to produce accurate text in images, Dall-E 3 claims to do so. This means users could produce things like brand logos, graphic ads with text for marketing campaigns, and other assets that combine text with images for work and personal use.
OpenAI says that images created with Dall-E will be free to use for commercial purposes without requiring permission to reprint or sell. This means you can use these images for work without worrying about legal issues.
Dall-E 3 will be rolled out to ChatGPT Plus and Enterprise users in a few weeks. OpenAI said that it will also allow artists to opt their art out of future versions of text-to-image AI models.
TOGETHER WITH PASSAGE BY 1PASSWORD
Protect your business and users from AI hackers by going Password-less
Passkeys protect your business and customers, while also driving business results
With AI challenging cyber security norms, traditional passwords are less secure than ever. AI can guess your password with 93% accuracy just by listening to your keystrokes and can crack 51% of common passwords in a minute.
That’s why the future of authentication is passwordless. Passkeys are a passwordless solution that leverages public-private key cryptography and device bound biometrics to create a 2x faster login with a 4x success rate. With passkeys, you get a better user experience and better security.
Google, Amazon, TikTok, and Whatsapp are just a few of the leading tech companies offering faster and more secure sign-in with passkeys. With Passage by 1Password, your business can bypass all the complexity of building an in-house solution, and implement passkey authentication in minutes.
Google is developing an AI-powered microscope to help doctors spot cancer.
Google has prototyped the Augmented Reality Microscope (ARM) with AI upgrades for real-time visual indications to help in categorizing samples and detecting cancer cells or infections in partnership with the Department of Defense. The Defense Innovation Unit intends to provide military ARMs at a cost of $90,000-$100,000.
DeepMind’s New AI Predict Genetic Diseases
DeepMind unveiled AlphaMissense, an AI model that forecasts the likelihood of missense variations leading to genetic illnesses with 90% accuracy. Missense mutations are single-letter genetic code changes that have the potential to cause illness. The algorithm discovered abnormal amino acid sequences in people with 71 million potential variations and assigned a "pathogenicity score" to each. This technique has the potential to speed up genetic illness research, but it is not a stand-alone diagnostic. Read more
AI Accelerates Ability to Program Biology Like Software
AI is transforming synthetic biology by allowing scientists to program live creatures in the same way that software is programmed. The discipline, which influences medical science, pharmaceuticals, food science, agriculture, and climate research, permits proteins to be repurposed for new roles. Read more
📈 Feature your product in the world’s biggest AI newsletter
Superhuman is the world’s biggest and fastest-growing AI newsletter with 400,000+ readers working at companies like Apple, Meta, Amazon, Google, Microsoft, and many more. Companies like Masterworks, Brave, SEMRush, and 1Password feature their products in Superhuman. Main ads are sold out for the next 6 weeks. You can book future ad spots here.
BE SMART: GO PASSWORD LESS
AREN'T HUMANS SMART?
TheRachele of AI
See more comments