The latest development, disclosed in an official blog post by Bing, includes the addition of voice chat for desktop users. The feature is already available for users who use the platform through mobile.

We know many of you love using voice input for a chat on Mobile. It’s now also available on desktop by clicking on the microphone icon in the Bing Chat box, the blog post read.

Access to the feature is facilitated via a microphone icon that desktop users with functional microphones can now select to begin speaking on the platform. According to Microsoft, the new feature enhances the user’s experience with the chatbot.

Notably, the platform offers multilingual support for an array of languages, including English, Japanese, French, German, and Mandarin, with additional languages to be added shortly. Additionally, text-to-speech technology is available, enabling the chatbot to answer user inquiries verbally in its unique voice.

Bing Chat has been receiving frequent updates, some every week, to enhance its usability. The latest update was released in conjunction with Microsoft Build. On June 2, Microsoft extended Bing Chat’s capacity to 30 turns per chat and 300 chats per day. Furthermore, Bing Image Creator’s support was expanded to all modes, and travel inquiries were optimized to provide more visual outcomes.

As a new development, the Vivaldi Browser has implemented a feature that disguises itself as Microsoft Edge on Android devices. The move is aimed at circumventing the browsing restrictions set by Microsoft on Bing Chat. Although the service is designed to be accessible only via Microsoft Edge, it is noteworthy that the browser is built on the same codebase as other Chromium-based browsers, such as Google Chrome, Brave Browser, and Vivaldi.

Via Microsoft Bing

As AI capabilities become increasingly integrated into products across the world, adhering to a bold and responsible framework will be even more critical, a Google blog post reads.

Cybersecurity executives from Google, Royal Hansen, and Phil Venables, wrote in a blog post that making AI models secure by default is a crucial aspect of implementing such models, and the introduction of the SAIF marks a significant first step towards achieving that goal.

As per the search giant, the SAIF system has been created to reduce the risks inherent in AI systems. Such risks include but are not limited to theft of the model, poisoning of training data, the injection of malicious inputs via prompt injection, and unauthorized access to confidential information present in the training data.

Google has asserted that SAIF can provide support to firms in protecting their neural network’s code and training dataset from theft. Additionally, it is highly effective in blocking a variety of other threats. The framework is primarily based on six distinct collections of best practices designed to help enterprises upgrade their AI security operations in different areas.

The Secure AI framework proposes six core ideas to enhance AI system security for organizations. These ideas include implementing existing security controls to new AI systems by adopting measures such as data encryption, expanding threat intelligence research to address AI-specific threats, leveraging automation in cyber defenses to react to anomalous activity quickly, conducting regular security reviews of AI models, continuously testing AI systems, and setting up an AI risk-aware team to mitigate business risks.

In the meantime, Microsoft is endeavoring to address the anxieties of the general public regarding the potential damage and abuse that could arise from unbridled AI deployment, which has prompted the company to impart its pledge of transparency through the Customer Commitments principle. Anthony Cook, the Corporate Vice President and Deputy General Counsel of Microsoft has unveiled three innovative AI Customer Commitments to allay concerns and foster a culture of trust in the organization.

Via CSO Online

The organization has authenticated a New York Times article that disclosed the employees being shown the preview of their new products, comprising ChatGPT-inspired chatbots intended for Messenger and WhatsApp. The said bots possess the capability of holding conversations using distinct personalities.

The meeting was held at the company’s headquarters situated in Menlo Park, which was streamed live to its global branches. Significant inputs were shared by the CTO, Andrew Bosworth, CPO, Chris Cox, and CEO and Founder, Mark Zuckerberg.

Interestingly, while Facebook’s parent company has been at the forefront of exploring generative AI technologies, it has been somewhat sluggish in implementing these innovations into its offerings.

At the meeting, Zuckerberg presented many AI technologies in different stages of development, with some intended for internal utilization but the majority designed for consumer use.

As an illustration, one of the updates in the pipeline would enable patrons to leverage a text command to edit their pictures and exchange them through Instagram Stories. Meanwhile, a separate development involves integrating AI agents with varying traits and features to provide assistance or amusement. The latter is presently targeted at employment in Messenger and WhatsApp.

Furthermore, the organization is organizing an in-house hackathon in July that concentrates on generative AI. The CEO has also recommended disseminating research findings and collaborating on AI technologies with the open-source community.

Meta, in February, underwent a reorganization of its AI initiatives to increase the implementation of the technology into its products, in line with its goals. In contrast, some companies have taken a more aggressive approach when incorporating generative AI tools into their products. For example, Microsoft has included AI in all its superior products, such as Office, Windows, and Bing. Moreover, a direct competitor of Meta, Snapchat, has launched its Snap AI chatbot.

Via The Guardian

The blog post explains that large language models (LLMs) like Bard are primarily prediction engines. When given a prompt, they generate responses by anticipating the next words in a sentence. While this makes them excellent at writing emails and essays, their performance in software development tasks has been somewhat error-prone.

Although code-generating models like GitHub’s Copilot and Amazon’s CodeWhisperer exist, they are not designed for general-purpose use. Unlike Bard and similar models such as ChatGPT, which were trained on a wide range of text samples, including web content, e-books, and other resources, code-generating models were mostly trained on code samples.

To address the limitations of general LLMs in coding and mathematics, Google developed implicit code execution, enabling Bard to write and execute its own code. The latest version of Bard identifies prompts that would benefit from logical code, generates the code internally, tests it, and incorporates the result into its response, aiming for greater accuracy.

According to internal benchmarking, Google claims that the new Bard has shown a 30% improvement in responding to “computation-based” word and math problems compared to the previous release. However, these claims will need to be verified through independent testing.

When Bard was initially launched earlier this year, it faced criticism and unfavorable comparisons to competitors such as Bing Chat and ChatGPT. The rollout experienced a setback when a Google ad featuring a wrong answer from Bard briefly caused an 8% drop in the company’s stock.

To address these issues and regain ground, Google has introduced implicit code generation and other enhancements, including support for new languages, multimodal queries, and image generation.

However, it remains to be seen whether these improvements will be sufficient to compete with leading generative AI chatbots in the field.

Via: TechCrunch

In a press authored by Microsoft’s corporate vp and deputy general counsel Anthony Cook, the company offers up three new AI Customer Commitments intended to offer additional levels of “guardrails” and transparency for its customers as they go about using AI today and in the future.

First, we will share what we are learning about developing and deploying AI responsibly and assist you in learning how to do the same. Microsoft has been on a responsible AI journey since 2017, harnessing the skills of nearly 350 engineers, lawyers and policy experts dedicated to implementing a robust governance process that guides the design, development and deployment of AI in safe, secure and transparent ways.

This includes sharing expertise via key documents and publishing’s as they become available. For Microsoft’s part, it will begin its shared documentation process with its Responsible AI Standard, AI Impact Assessment Template, AI Impact Assessment Guide, Transparency Notes. 

Second, we are creating an AI Assurance Program to help you ensure that the AI applications you deploy on our platforms meet the legal and regulatory requirements for responsible AI.

The program includes regulatory engagement support, risk framework implementation, customer councils, and regulatory advocacy.

Third, we will support you as you implement your own AI systems responsibly, and we will develop responsible AI programs for our partner ecosystem.

Microsoft intends to offer dedicated resources in the form of AI legal and regulatory experts around the world as well as a comprehensive partner network to aid in launching and leveraging new AI platforms.

Admittedly, Microsoft understands that its proposed AI Customer Commitments are only the start of what could be a long AI journey for some businesses but at this moment in time, these commitments need to be considered and adopted by the sector at large.

The tech giant is slated to unveil a new feature on Wednesday that will allow Azure Government customers to leverage the prowess of two of OpenAI’s expansive language models, namely, GPT-4, the start-up’s latest and most advanced model, and GPT-3, an earlier variant.

The forthcoming release is likely to refrain from explicitly mentioning the US government agencies that are anticipated to adopt the technology, however, Bloomberg suggests that the US Defense Department, the Energy Department, and NASA could potentially be included, given that they are federal government clients that have utilized Microsoft’s Azure Government solution.

The move provides government entities access to two of OpenAI’s large language models, namely the most recent and potent GPT-4 and its antecedent GPT-3, through Microsoft’s Azure OpenAI service. It has been confirmed that the Defense Technical Information Center (DTIC), an arm of the Defense Department that specializes in assembling and distributing military discoveries, will conduct trials employing the GPT models via Microsoft’s latest offering.

According to a recent blog post reviewed by Bloomberg, Microsoft’s government clients can leverage OpenAI’s LLMs for a vast array of tasks such as answering crucial research queries, computer code generation, and field report summarization. Azure Global CTO William Chappell revealed this information in the post. Nevertheless, the company’s representative noted that these agencies could not access ChatGPT directly, though they could utilize OpenAI’s models through a chat-like interface.

Microsoft has designated a segregated commercial cloud computing space to accommodate the OpenAI models. The area is distinct from the cloud utilized by Azure Government customers, which strictly observes various security and data compliance regulations. As per Chappell’s statement, the data transmitted to the models remains within the Azure OpenAI service since Microsoft hosts them in its Azure infrastructure. Moreover, it was also mentioned that the AI models won’t be trained using data from Azure Government customers.

Via Bloomberg

According to Microsoft via an email:

Four months after Microsoft introduced the new Bing and began to transform the expectations and experiences of Search, people have engaged in more than half a billion chats, created more than 200 million images with Bing Image Creator, and usage of the Bing mobile app has increased 8x since launch.

The company further cites that despite being able to hit these numbers, users are yet to leverage the tool’s full potential, which is why it’s launching the Bing Chat campaign. “Microsoft’s campaign aims to address this challenge by sparking creativity and curiosity to highlight the new possibilities of asking more fun and compelling questions,” Microsoft told OnMSFT.

The campaign will essentially help users learn how to give prompts, opening them up to a new world with endless possibilities. Here are a couple of illustrations of how much you can exploit the tool’s capabilities to get the exact response you might be looking for:

Microsoft should roll out additional campaign assets across digital, social, and OOH platforms in the next few weeks. You can check out more prompts here.

In the past, we’ve seen reports surface online with several users citing instances where the chatbot had hallucinated or outrightly given wrong responses. Perhaps, this campaign will help users learn how to use the tool properly, ultimately helping them make the most out of it.

How do you make your prompts while using Bing Chat? Share your experience with us in the comments.

Musk, who recently visited China, indicated that his conversations with senior Chinese officials were productive. Although he refrained from divulging explicit details, he emphasized the importance of oversight and regulation for the development of AI.

“Something that is worth noting is that on my recent trip to China, with the senior leadership there, we had, I think, some very productive discussions on artificial intelligence risks, and the need for some oversight or regulation,” Musk said.

“And my understanding from those conversations is that China will be initiating AI regulation in China.”

There has been a surge of activity among Chinese corporations eager to create AI-driven services that can replicate human conversation ever since OpenAI, a San Francisco-based company, introduced ChatGPT in November.

However, the swift progress made in this field has aroused concerns worldwide regarding the risk of disinformation and the possibility of unethical use of this technology.

Although Musk did not provide specific details, his remarks regarding his interactions in China are speculated to allude to comprehensive preliminary legislation, which mandates that any newly developed AI products must undergo a security evaluation before deployment and adhere to fundamental core socialist values.

Several prominent figures in the field of AI have authored an open letter warning about the potential for AI to lead to human extinction. Signatories to the letter include Sam Altman, CEO of OpenAI. However, recent reports have suggested that the prediction is likely overblown.

Via Reuters

According to a recent blog post by Alphabet, Gmail users on mobile devices can anticipate the integration of a “top results” section, which will enhance search accuracy when retrieving past messages or attachments through the Gmail app.

Updates are being rolled out to the mobile apps for both Google Workspace and personal Google accounts, which will include new machine-learning models. Among the changes, users will notice a new section called “Top results” appearing at the top of their search results.

The new AI-powered feature will consider various factors, including the user’s recent emails, to present the most relevant search results.

“When searching in Gmail, machine learning models will use the search term, most recent emails, and other relevant factors to show you the results that best match your search query. These results will now appear at the top of the list in a dedicated section, followed by all results sorted by recency,” said Google in a blog post.

As stated in Google’s official blog, the “top results” feature has been in high demand and will soon be accessible to all Gmail users on their mobile devices within the next two weeks. However, It is currently uncertain if the function will also be integrated into the web-based edition of Gmail.

Google has exhibited a strong dedication to integrating AI into its offerings, as evidenced by the launch of its latest feature. The tech giant rolled out multiple AI-focused services and features during its I/O developer conference held in May.

The release of OpenAI’s ChatGPT, a well-known generative AI platform, and Microsoft Bing prompted the proliferation of advancements and substantial financial investments in the field. Consequently, Google initiated efforts to keep pace with their popularity, culminating in the release of its AI chatbot Bard. 

Via PCMag

Altman was questioned about the possibility of his company setting up a local office during his visit to Microsoft Corp’s research and development center in Israel. As per Microsoft’s official statement, he conveyed that while his company favors working from a single location, various investment options in Israel are currently being evaluated.

Moreover, Altman reassured that despite the increasing automation trend, he dismisses any concern about the potential impact of AI on employment. He emphasized that there will always be work opportunities for human beings and asserted that the jobs of the future, almost a century from now, will be remarkably different from those of the present.

During his speech at Tel Aviv University, the OpenAI CEO lauded the exceptional talent pool in Israel and expressed his conviction that the local tech ecosystem will undoubtedly play a pivotal role in the ongoing AI revolution.

The widespread adoption of generative AI, particularly after the introduction of Microsoft-sponsored ChatGPT by OpenAI last year, has spurred lawmakers worldwide to consider safety concerns surrounding this technology. The European Union is leading the charge with its proposed AI Act, slated to be ratified later in the year.

Meanwhile, the United States appears to be more inclined towards modifying existing regulations than creating entirely new ones for AI. British policymakers are also keen to ensure that any legislation on the matter does not inadvertently stifle innovation.

Following his visit to Israel, Altman has scheduled trips to Jordan, Qatar, the United Arab Emirates, India, and South Korea, all within the next week.

Via Reuters