Generative AI Across Social Media Platforms (2023)

Generative AI is swiftly asserting its dominance online, with its influence becoming particularly influential on social media. Defined by Social Media Today as “a type of artificial intelligence (AI) that uses machine learning algorithms to create new and original content like images, videos, text and audio”. Generative AI’s capacity is arguably producing original content that rivals human-created material. It is not merely a trending topic—it represents a paradigm shift in how content is generated, shared, and engaged with, across various social media platforms.

AI-powered tools streamline social media management by automating tasks, thereby enabling the creation of more targeted and effective content strategies. Content creation is a critical component of social media success, requiring a continuous stream of fresh, appealing, and relevant material to capture and maintain user attention. In such an environment, generative AI emerges as an invaluable ally for creators, marketers, and social media managers.?

Let’s take a look at how various social media platforms are using Generative AI in 2023:

Meta: Facebook & Instagram

AI Stickers

Meta has introduced a new range of AI stickers, enabling users to generate personalised stickers for use in chats and stories across various platforms, including WhatsApp, Messenger, Instagram, and Facebook Stories. This feature, using technology from Llama 2 and Emu, translates text prompts into a selection of unique stickers.

Source: Meta Newsroom?

Image Editing Tools on Instagram

• Restyle- This tool allows users to transform their images by applying visual styles based on text descriptors, ranging from simple terms like “watercolor” to more complex prompts like “collage from magazines and newspapers, torn edges”.
• Backdrop- Users can alter the background or scene of their images using prompts, integrating the primary subject with the newly described backdrop.

Source: Meta Newsroom?

Conversational AI Assistant (Meta AI)

Meta AI, now in beta, stands as a conversational assistant accessible via WhatsApp, Messenger, and Instagram. There are also plans for expansion to Ray-Ban Meta smart glasses and Quest 3. Powered by advanced AI models, Meta AI offers real-time information, image generation capabilities, and interactive experiences tailored to users’ needs and inquiries.

Source: Meta Newsroom?

AI Characters

Meta is rolling out AI characters, each with distinct personalities, interests, and backstories. Some of these characters are brought to life by celebrities and influencers, providing users with a wide array of interactive experiences. These AI characters are intended to add a layer of familiarity and entertainment to user interactions. Find out more here.

Source: Meta Newsroom?

Generative AI for Advertisers

Meta has unveiled a series of generative AI features specifically designed to enhance the capabilities of advertisers on its platforms. These tools aim to streamline the creative process, offering unique ways to modify and optimise ad content.

• Customisable Creative Assets- Advertisers can now generate a variety of backgrounds to alter the appearance of their product images.?
• Image Expansion- This feature addresses the need for diverse aspect ratios across different Meta platforms such as Feed or Reels. Image expansion allows advertisers to adjust their visual assets, including images and videos, to suit various display requirements.
• Text Variations in Ads- The generative AI can produce up to six different text variations from the advertiser’s original copy, emphasising specific keywords or phrases.?

Source: TechCrunch

Privacy

Privacy is a top priority for Meta as it develops its generative AI features. The company says that it is committed to building these features responsibly, with a focus on ensuring user data is protected.?

• Model Training and User Privacy-Release of Llama 2– Meta released its large language model for academic and commercial use, ensuring that Meta user data was not included in its training.Building Consumer Generative AI Features- These are built upon the Llama 2 model, with additional privacy safeguards and integrity mitigations integrated.

• Data Sources and Usage-Combination of Sources- A variety of data sources were used to train the generative AI models.Exclusion of Private Information- Steps were taken to filter out websites that commonly share personal information and to exclude private posts and messages from platforms.

• Handling of Information Shared with AI Features-Personalised Responses- Information shared in a chat may be used to provide more personalised and relevant responses within that conversation.
• Sharing with Trusted Partners- Certain queries may be shared with trusted partners, such as search providers.
• Limiting Exposure of Private Information- Measures are in place to limit the possibility of private information being shared in AI responses to other users.
• Review of Interactions- Both automated technology and human reviewers are employed to ensure the safety and performance of the AI models.

• User Control-Deletion Commands- Users have the ability to delete information shared in AI chats.
• Separation of Account Information- Using Meta’s generative AI features does not link WhatsApp account information to accounts on other Meta platforms.

• In-Product Transparency Tools- Tools are available to help users identify content created by AI features, including visual watermarks on AI-generated images.

To see the full outline of the steps Meta has taken to ensure privacy and transparency, click here.

Responsible AI Development

Meta says that they are dedicated to responsible AI practices, ensuring the safe performance of their products through rigorous testing. They are actively engaging with policymakers, academic experts, civil society, and industry peers to promote the responsible utilisation of this technology. The gradual rollout of these features, starting with beta versions of the AIs, allows Meta to continuously refine and enhance them based on real-world usage.

• Image Creation Transparency-Visible Markers- Images created or edited by Meta AI features bear visible markers indicating AI involvement.
• Industry Standards- Meta advocate for industry-wide standards for labelling AI-generated content and actively participate in relevant forums.

• Misinformation Mitigation-Content Matching and Adaptation- Meta employs AI to identify near-duplicates of fact-checked content and quickly respond to emerging harmful content types.
• Generative AI’s Potential- Preliminary tests indicate that large language models trained on Meta’s Community Standards could enhance their policy enforcement capabilities.

To see the full outline of the steps taken to ensure generative AI features are built responsibly, click here.

Content Ranking

Meta’s AI systems work to predict the potential value of content to users, aiming to present the most relevant posts promptly. A combination of behaviour-based and feedback-based predictions are utilised, with the integration of wide-ranging signals to refine content recommendations. And, to promote transparency, Meta has introduced “system cards,” offering insights into AI operations and user customisation options across various content sections.

Meta emphasises the power users have in shaping their platform experience, encouraging the utilisation of available tools to tailor content visibility. Centralised locations for customisation controls have been established on both Facebook and Instagram. Features for content preference indications are being tested, alongside existing options to adjust content visibility and ranking.

Source: Meta Newsroom

Community Forum

Meta is strategically utilising generative AI to foster inclusivity and diversity in decision-making through the establishment of Community Forums. By integrating deliberative polling methodologies, these forums present a platform for users to engage with complex issues, such as the development and application of generative AI technologies. Participants are provided with extensive educational resources, enabling them to make informed contributions during discussions.

This initiative aims to gather a wide range of perspectives, ensuring that the values embedded in AI systems reflect the diversity of the global community.

By opening up the development process and inviting diverse viewpoints, Meta aims to lay the groundwork for responsible and ethical AI systems.

Source: Meta Newsroom

AI in Europe

Meta’s move to introduce an AI off-switch in Europe, aligning with the Digital Services Act (DSA), is essential for compliance with the EU’s revamped digital regulations, ensuring users have the autonomy to switch off AI-driven personalisation in content feeds on platforms like Facebook and Instagram. The DSA’s main objective is to give users the option to opt for a non-algorithmic feed, ultimately promoting user autonomy and choice.

Source: TechCrunch

Instagram Chatbot

Instagram has an AI chatbot feature in DMs. This new feature can be likened to Snapchat’s My AI tool, bringing AI capabilities directly into direct messaging. Users can interact with the AI tool, asking questions or seeking advice on crafting more impactful messages.

The process is initiated by typing @ai in the chat field, redirecting users to the AI chatbot.?

While the functional impact of the Instagram AI chatbot may not be groundbreaking, its introduction is a strategic move to acclimate users to AI interactions. This is a step towards more innovative and potentially revolutionary applications of generative AI in the future, such as the creation of VR worlds.?

Source: Social Media Today

Future Developments

• AI for Businesses and Creators- Meta is extending AI capabilities to businesses and creators, enabling them to craft AIs that resonate with their brand values and enhance customer interactions. This move is part of a broader initiative to democratise AI creation, making these tools accessible to a wider audience.
• AI Studio and Developer Support- The introduction of AI studio lays the groundwork for future innovations, empowering both coders and non-coders to develop their own AI experiences. Meta plans to open up APIs for third-party AI development, starting with Messenger and eventually expanding to WhatsApp.
• Ad Copy- Meta is working on additional AI-driven tools to generate ad copy that highlights key selling points and backgrounds with thematic relevance.?
• Customer Interactions- Businesses will soon be able to leverage AI for customer interactions on WhatsApp and Messenger, covering various aspects such as e-commerce, engagement, and support.

……………………………………………………………………………………………….

LinkedIn

Responsible AI Principles

LinkedIn is embracing transformative advances in AI. AI has been a longstanding component of LinkedIn’s toolkit, playing a crucial role in enhancing professional experiences for its members by fostering connections, boosting productivity, and aiding in career success.?

Drawing inspiration and alignment from Microsoft’s leadership in Responsible AI, LinkedIn announced that they are adhering to a set of 5 Responsible AI Principles to guide its endeavours:

1. Advance Economic Opportunity- AI is viewed as a tool to amplify the success and productivity of its members, helping to realise their professional aspirations.
2. Uphold Trust- Proactive measures are taken to mitigate potential risks associated with AI.
3. Promote Fairness and Inclusion- LinkedIn is dedicated to ensuring that the benefits of AI are distributed equitably among all its members, actively working to prevent and rectify unfair biases that may arise.
4. Provide Transparency- LinkedIn strives to demystify AI, offering clear and simple explanations to illuminate how the technology affects users.
5. Embrace Accountability- LinkedIn upholds robust AI governance practices, encompassing thorough assessments of potential harms, fitness for purpose, human oversight, and accountability.?

Source: LinkedIn Blog

?

LinkedIn Premium Accounts

For premium account users, LinkedIn is adding AI-powered features, including prompts to help you write your profile and create content.?Watch this video for a summary of the AI-powered LinkedIn Premium account experience:

Filter with AI

LinkedIn is making significant strides in integrating generative AI into its platform, particularly for those utilising Sales Navigator. The introduction of the “Filter with AI” table marks a substantial leap forward. It presents users with a means to refine their searches through prompts and conversational queries. This makes it simpler to pinpoint precise leads and significantly broadens the scope of research.

Source: Social Media Today?

Account IQ

The “Account IQ” feature, a generative AI tool crafted to provide succinct summaries of information regarding potential contacts. This feature aggregates key details such as recent news, financial data, and information on key team members, all collated and presented by LinkedIn’s advanced AI system.?

Source: Social Media Today

Account Hub

LinkedIn has introduced an account prioritisation feature that intelligently pinpoints vital contacts for sales pitches, utilising a multitude of signals. This tool aims to simplify the complexity of managing daily changes such as new hires or a customer leaving your account. By uploading a client list or book of business, users can use the Account Hub to prioritise based on aspects like growth potential, buyer intent, or recent activities like new hires.?

Source: Social Media Today

Personas

The “Personas” feature is another significant addition. It enables marketers to craft templates for their ideal lead candidates. Sales Navigator then uses to surface the right contacts based on specified parameters like function, seniority, and geography.

Source: Social Media Today

Recruitment

On the Recruiter side, the “Likelihood of Interest” feature has been introduced. This will highlight members more inclined to engage and respond to outreach efforts, drawing on a variety of activity-based signals. This AI-driven functionality aggregates real-time interest signals, identifying candidates who are either highly or moderately likely to be interested in a job.

Source: Social Media Today

The new Recruiter interface is designed to be intuitive. It employs natural language commands that allow users to easily initiate searches, such as “Find me a senior growth marketing leader.” A ChatGPT-like prompt window facilitates deeper inquiries into specific elements, providing personalised nudges based on unique member insights. This could range from suggestions to expand the search location, to adding in-demand skills, or advising on the most suitable work arrangement for a particular role.

LinkedIn is also introducing CRM Connect, aiming to integrate LinkedIn Recruiter and CRM systems for a more unified experience. This feature is currently in trial with a select group of partner businesses.

Source: Social Media Today

Predictive Audiences

LinkedIn’s Campaign Manager has incorporated predictive audiences into its targeting tools, powered by the platform’s advanced predictive AI. This enables marketers to identify and connect with high-intent audiences efficiently at scale.?

Predictive audiences leverage their user base and engagement levels to enhance your first-party data, allowing for scaled targeting through predictive AI modelling. This ensures that your campaigns are more likely to connect with individuals who are ready to take action.?

Source: LinkedIn Blog

AI Prompts for Ad Creation

The new AI feature in Campaign Manager is designed to aid users in crafting more effective LinkedIn advertisements. The tool provides multiple variations of ad copy and headlines. It does this by utilising data from the company page, campaign insights, and the power of AI.

Source: Social Media Today

LinkedIn Learning

LinkedIn is incorporating generative AI into LinkedIn Learning. A new AI-powered chatbot experience will offer real-time advice and tailored content recommendations. These will be personalised based on job title, career goals, and followed skills. This initiative aims to provide users with a virtual career coach, accessible at any time.

LinkedIn is launching the first phase of its new AI chatbot for LinkedIn Learning, starting with leadership and management skills. The chatbot will engage users in interactive dialogues. It will pose clarifying questions to better understand their specific situations before providing advice and examples. LinkedIn plans to expand and further tailor this chatbot experience, potentially including responses from specific industry experts.

Source: Social Media Today

?

……………………………………………………………………………………………….

X (formerly Twitter)

AI Investments and Features

Twitter could potentially be the next major platform to integrate generative AI. Elon Musk reportedly acquired thousands of GPUs for an initiative to develop proprietary AI tools.

The potential integration of an AI model built entirely on tweets is intriguing. Athough it raises questions about the tone and nature of the generated content. The application of this technology could range from enhanced content generation to customer service tools in Twitter DMs. And, even visual prompts responding to tweet prompts.

With the recruitment of former DeepMind engineers and significant investment in the necessary hardware, it seems that Twitter’s venture into generative AI could materialise in the near future. The exact application and impact of these tools remain to be seen. However, they have the potential to reshape user interactions and content generation on the platform.

Source: Social Media Today

?

……………………………………………………………………………………………….

TikTok

AI Ad Script Generator

TikTok has introduced the ‘Script Generator’, an AI-powered tool located within its Creative Center. Users input details about their industry, product, and related keywords. Then, they choose their desired video length. The tool is prompted to generate a selection of sample scripts complete with audio and visual cues.

Source: Social Media Today

Labels for AI-Generated Content

• Implementation of AI Labels- TikTok is enforcing rules that necessitate the tagging of AI-generated content. The aim is to mitigate the potential risks associated with deepfakes and misleading information.
• The Need for Transparency- As AI-generated content becomes more prevalent, transparency and user awareness are crucial. Particularly, during critical events like conflicts, elections, and natural disasters where misinformation can have significant consequences.
• Potential for Expansion- The conversation around tagging and transparency could extend to text-based content generated by tools like ChatGPT. This emphasises the need to identify and manage AI-driven interactions across different media types.

Source: Social Media Today

?

……………………………………………………………………………………………….

Conclusion

Reflecting on the above information, it’s evident that we are witnessing a transformative era in social media. Platforms like Meta, LinkedIn, X (Twitter), TikTok, and others are rapidly adopting AI technologies to enhance user experience, improve content relevance, and streamline operations. However, this integration comes with its share of challenges and considerations.

What AI tools and features (if any) are you currently using or plan on using to manage your social media activities?